00:46:38 Freya Anderson, she/her: Reacted to "Hello, fellow humans..." with ❤️ 00:46:44 Houcine Al-Najaar: Hello everyone. I'm Husayn from France 00:46:57 Irene Tobón Restrepo: Buenos días desde Colombia! 00:47:09 Andrew Lopez: New London, CT, in the house, cheers! 00:47:17 John Shafer: Hello from foggy San Francisco, a few blocks away from the physical Internet Archive. 00:47:29 Duncan Hall: Hi all! I’m Duncan Hall and I work with the Patron Services team at the Internet Archive. I’ll be sharing links & information during today’s session. 00:47:29 Andrea - VPL (she/her): Hi from Vancouver, Canada 00:47:39 Alana Krause: Reacted to "Hi all! I’m Duncan H..." with 👍🏻 00:47:42 MaryBeth Matthews: Greetings from Annapolis, MD! 00:47:43 Clifford Anderson: Hello from Fishers Island, NY (and remote at Vanderbilt U)! 00:47:43 Luis Villa (he/him): Creating a slide deck without using GPT image generation is now illegal 00:47:45 Mark Wilson: Greetings from Whittier, California! 00:47:48 charles blass: whassupp from ztown zürich ✌🏼🌊🪐💫 00:47:49 rob: here in the RGV valley in texas... this is Roberto 00:47:55 Ema Mezihoráková: Reacted to "Hi all! I’m Duncan H..." with 👍🏻 00:47:56 Nate Angell (he/him): Reacted to "Creating a slide dec..." with 🤪 00:48:06 Maya: Namaste from Pune, India..Maya Dodd here 00:48:13 Alan Levine: Hello from a human in Moose Jaw, SK. No ChatGPT is used in my chats. 00:48:21 Tyler Dool: Howdy, from Dallas Texas. 00:48:28 M.Mubashir Farooq: Hi from malaysia 00:48:28 Judith Sebesta: Reacted to "Hello from a human i..." with 😀 00:48:32 raquel.donahue: Hi from Houston Community College 00:48:32 Sam Eddington: Hi from Salisbury, Maryland! 00:48:33 Belén: Hola! Hi from Chile! 00:48:33 Dan Zajdband: Hello from Madrid, Spain 00:48:35 Charles A.: Hello form MD 00:48:35 Luis Villa (he/him): [and hello, also from San Francisco!] 00:48:35 Roxanne: Hello from Alberta, Canada 🙂 00:48:36 Duncan Hall: Captions are available for today’s discussion by turning on Zoom’s Live Transcript feature. Our chat is open for today’s discussion. Please keep your comments respectful and on topic. Today’s session is being recorded. All registrants will receive an email with a link to the recording and the resources shared in chat. 00:48:37 mdagostino: Hi from Philly 00:48:37 Laura Erickson: Good morning from San Francisco! 00:48:37 Mahi Jain: Namaste from India 00:48:38 Patrick_F: Motor city says hey 00:48:39 Maria Souliotis: Osiyo (Hello) from Tahlequah, OK, the capital of the Cherokee Nation. 00:48:39 Charlie: Hi, joining from very sleepy Tokyo, Charlie here. 00:48:39 Jessi Haley (she/her): Hello from Healdsburg, CA! 00:48:39 Delena: Hello from Chicago! 00:48:40 Violet Price: Hello from New York 00:48:40 Arnold Rafael: Hello from SF, CA, USA :) 00:48:41 Cable Green: Greetings All: Cable Green, Director of Open Knowledge at Creative Commons (Olympia, WA, US) 00:48:41 Carolyn Holcroft: Hello from San Jose, California! 00:48:43 charles blass: http://link.bar/cb 00:48:45 Lenora Gobert: Lenora Gobert from New Orleans, Louisiana. Hi! 00:48:46 Ali Barış Öndül: Hello, Barış from Izmir, Turkey <3 00:48:46 Duncan Hall: Our chat is open for today’s discussion. Say hello! Where are you joining in from today? 00:48:47 Ludovic Dubost : Hello, from Paris, France 00:48:47 Tamiko Thiel: Hi, Tamiko Thiel here checking in from Munich, Germany. 00:48:47 Jacob Rogers: Hello from Berkeley, California! 00:48:49 Kimi Taira (she/her): Hello from Los Angeles! 00:48:50 Matéo Munoz: Hello from France 00:48:51 Mike Trizna (he/him): Greetings from Washington, DC! 00:48:51 Larry Dieterich: Greetings from Davis California 00:48:51 HW Yaboah: Hello fromOC, CA 00:48:51 Asako Yoshida: Hi! from Winnipeg, MB, in Central Canada! 00:48:52 Mark Graham: Greetings from Half Moon Bay, CA 00:48:52 Alana Krause: Reacted to "Greetings All: Cable..." with 👏🏻 00:48:52 Robson Silva: Hi. Robson, from Brazil. 00:48:53 DINA_ Pend.kimia: HI, from Indonesia 00:48:53 Joshua Fontany: Hello from coastal Oregon. 00:48:58 conail: Reacted to "Hello from Alberta, ..." with 😎 00:48:58 Adam Goldberg: hello from NYC 00:48:59 Joshua Fontany: Reacted to "HI, from Indonesia" with 🙏 00:48:59 Carrie Marlowe: Two of us in Rahway, NJ 00:48:59 claudia meglin: nevada city CA 00:49:01 Ern Bieman: Hello all! Ern Bieman from the Canadian Heritage Information Network. 00:49:01 Shafali: Hello I'm shafali:) a designer/researcher from bhopal, india 00:49:01 Jeff Glassman: Greetings from Long Beach California! 00:49:01 John & Sharon McFarland: Hello from Lakeland Florida 00:49:02 Nancy: Hello from NYC 00:49:03 Janice Banser: Hello from Burnaby, BC in Canada. 00:49:03 Ema Mezihoráková: hi from Berlin, Germany (I might leave early sadly) 00:49:05 Yael Eylat Van Essen: Hello from Prague 00:49:06 BZ: Hello from Portland OR. 00:49:07 Martha Rans: Martha Rans from Vangroovy BC 00:49:08 Jessica Tam: Hello from Oakland, CA 00:49:09 natan.lins: Hello Everyone...Natan from Brazil!! 00:49:12 Alan Levine: Hello Salisbury, MD! I remember going through there on my way to Ocean City 00:49:16 Laura Manrique: Hello! from Colombia! 00:49:18 Pete Kaminski: Good morning from San Diego, California. 🙂 00:49:21 Wigbert Boell: Hello from BERLIN, Germany 00:49:31 Fartein Hauan Nilsen: Greetings everyone! Anthropologist interested in the social uses of AI here. Currently located at UC Berkeley 00:49:33 Sergio Santamarina - UNPAZ: Hello everyone, Greetings from Argentina! 00:49:34 Duncan Hall: AI @ IA: https://blog.archive.org/tag/ai-research/ 00:49:35 Michael Pate: Hello from Sunny Central Florida 00:49:36 Giulia Taurino: Hi! Based in Boston, MA at Northeastern University 00:49:37 Michael Owen: Hello everyone Michael Owen from Brooklyn, NY 00:49:43 Jennifer Zimmer She/Her/Hers: Hi from little ol' Dowagiac, Michigan! We're just hoping it doesn't snow again today. 00:49:43 Imwen Eke: Hello from London. 00:50:04 Duncan Hall: Extracting Words Sung on 100 year-old 78rpm records: https://blog.archive.org/2023/04/12/aiia-extracting-words-sung-on-100-year-old-78rpm-records/ 00:50:08 Evo Heyning: Great to see you all, Evo Heyning here in Oakland, CA at Realitycraft, working on storytelling, music videos and immersive worlds + books 00:50:19 Amalyah Keshet: Hello all, from Jerusalem 00:50:21 Justin: Hello Justin from Tucson Arizona 00:50:26 Karen Traub: Greetings from Shutesbury Massachusetts 00:50:35 Wigbert Boell: Reacted to "Extracting Words Sun..." with 👍🏼 00:50:44 Pete Kaminski: https://github.com/tesseract-ocr/tesseract 00:51:04 Nate Angell (he/him): Greetings from Nch'i-Wána lands, traditional home to the Clackamas, Cowlitz, Grand Ronde, and Siletz, also known as Portland, Oregon, USA 00:51:07 Tanya V: Reacted to "Great to see you all..." with ❤️ 00:51:24 Duncan Hall: TV Third Eye: https://mastodon.archive.org/@tv 00:51:35 Olga Volgshtein: 😄 Hi from Russia 00:51:49 Wigbert Boell: Replying to "TV Third Eye: https:..." Mastodon! great! 00:51:56 RBallack: Vancouver, British Columbia. Isaac Asimov fan. He Saw it coming. 00:51:56 Stella Bastone: Not gonna lie — carbonating here over hearing Brewster Kahle speaking live since my students learn about him and his work in my course in Toronto, Canada. So grateful for this and for the entire session. 🙂 00:51:57 Shivam Agrahari: Hi everyone 00:52:11 Lila Bailey (Internet Archive, she/her): Reacted to "Not gonna lie — carb..." with ❤️ 00:52:14 Stephanie Beene: Reacted to "Not gonna lie — carb..." with 👆🏼 00:52:14 conail: Reacted to "Not gonna lie — carb..." with ❤️ 00:52:20 Duncan Hall: AI @ IA: https://blog.archive.org/tag/ai-research/ 00:52:27 Duncan Hall: Internet Archive weighs in on Artificial Intelligence at the Copyright Office: https://blog.archive.org/2023/04/28/internet-archive-weighs-in-on-artificial-intelligence-at-the-copyright-office/ 00:52:44 Sylvia: is there a link to the project you want volunteers for? 00:52:49 Laura Erickson: Reacted to "Not gonna lie — carb..." with ❤️ 00:52:50 Mark Graham: https://twitter.com/MarkGraham/status/1640510980714283009 00:53:10 Olga Volgshtein: https://blog.archive.org/2023/04/28/internet-archive-weighs-in-on-artificial-intelligence-at-the-copyright-office/ 00:54:06 Shivam Agrahari: does AI really affect our daily lives? 00:54:14 Cynthia Walker: Hi, I can't seem to hear, though the audio on my computer seems to be fine. Any tips? 00:54:19 Sylvia: Replying to "does AI really affec..." yes. it's already used in advertising 00:54:28 Marc-Antoine Parent: We hear, so probably on your side 00:54:29 Brewster Kahle: A task we would love help on: https://blog.archive.org/2023/04/24/ai-audio-challenge-audio-restoration-based-on-expert-examples/. And finding articles in scanned journal issues ( brewster@archive.org ) 00:54:32 Yajat Pal: hello wveryone 00:54:33 Aaron Miller: Replying to "Hi, I can't seem to ..." Try your Zoom preferences 00:54:35 Martha Rans: Interesting article from Canada:https://www.theglobeandmail.com/arts/art-and-architecture/article-as-ai-debate-swirls-artists-are-torn-between-embracing-it-and-trying/ 00:54:58 Yao Grace: Replying to "Hi, I can't seem to ..." Re-login 00:54:58 Pete Kaminski: @Cynthia, try changing your audio device from what it is to something else, and back. or, maybe drop and rejoin? 00:55:06 Mark Graham: This is very experimental: https://newsum.sawood-dev.us.archive.org (e.g. it does not yet filter out ads) (Summaries of top stories/day from some TV news channels) 00:55:10 Alana Krause: Replying to "Hi, I can't seem to ..." Change speaker in zoom or check sounds settings on computer. 00:55:11 conail: Replying to "Hi, I can't seem to ..." Switch to phone audio and back again. 00:55:15 O.P.A.: Replying to "Hi, I can't seem to ..." If audio continues to not work, try turning on closed captioning: maybe you can follow on with the subtitles if nothing else 00:55:36 Caitlin Olson: Replying to "is there a link to t..." Feel free to email me at caitlin@archive.org if you’re interested in being looped into the AI Audio Challenge Brewster posted 00:55:59 Sylvia: Reacted to "Feel free to email m..." with 👍 00:56:02 Yao Grace: Replying to "does AI really affec..." Yep, you thought you are talking to a human but a chatbot. 00:56:02 Cynthia Walker: Sounds working now. Thanks everyone! 00:56:14 Marc-Antoine Parent: Reacted to "Sounds working now. ..." with 👍 00:56:16 Caitlin Olson: Reacted to "Sounds working now. ..." with 🥳 00:56:33 Wigbert Boell: Reacted to "Yep, you thought you..." with 👍🏼 00:56:38 Yajat Pal: Reacted to "Sounds working now..." with 👍 00:56:40 conail: Reacted to "Feel free to email m..." with 👍 00:56:43 rob: Reacted to "Sounds working now..." with 👍 00:56:57 charles blass: what is the meaning of "openai commons" when "openai" is not "open" 00:58:10 Mark Graham: Please feel free to email me: mark@archive.org if you would like to connect about AI projects/experiments at the Internet Archive We love partnering! 00:58:20 charles blass: new exciting incentives to feed the machine's exponential appetite 00:58:30 Alan Levine: @ Charles “Hey ChatCPT, tell me how irony works!” 00:58:45 chris lange (he/him): Replying to "does AI really affec..." highly recommend checking out this revealing interview with Timnit Gebru on the misleading and major implications of AI: https://www.techwontsave.us/episode/151_dont_fall_for_the_ai_hype_w_timnit_gebru 00:58:46 charles blass: Reacted to "@ Charles “Hey ChatC…" with 😂 00:58:50 Stephanie Beene: Reacted to "@ Charles “Hey ChatC..." with 😂 00:58:56 conail: Reacted to "@ Charles “Hey ChatC..." with 😂 00:59:03 Stephanie Beene: Reacted to "Please feel free to ..." with 👋🏼 00:59:10 Marc-Antoine Parent: Reacted to "highly recommend che..." with 💯 00:59:32 Mark Wilson: Corporate chat bots will soon become lethal autonomous weapons. Corporate LAWs. 00:59:41 maria bustillos: Reacted to "Please feel free t..." with 👋🏼 00:59:41 Stephanie Beene: Replying to "does AI really affec..." Yes, it's already being used by Alexa and Siri and Google Nest Thermostat -- but more broadly, it's being implemented for city planning etc 00:59:59 Brewster Kahle: Thank you Lila. We hope people help find evidence for and against assertions in the collections. 01:00:14 rob: 2023 ai boomed... early with jan if im correct. 01:00:16 Brewster Kahle: And not just people but machines find... 01:00:17 Atticus Garrison: Chat bots as disinformation machines that churn out targeted disinformation clickbait articles are just a matter of time as well 01:00:26 Marc-Antoine Parent: Reacted to "Chat bots as disinfo..." with 💯 01:00:46 Andrew Padilla | Datacequia: Reacted to "Chat bots as disinfo..." with 💯 01:00:47 Stephanie Beene: @atticus, this is already happening 01:00:49 conail: Reacted to "This is very experim..." with 👆 01:00:51 R.G.Opus https://name.pn/ritabrata-ghosh: A great endeavor put forward in the part of Internet Archive. Thank You Lila! 01:01:15 Chris Freeland: As you have questions for our panelists, please drop them here in chat & we’ll work them into the conversation. 01:01:24 Stephanie Beene: https://www.popsci.com/technology/chatgpt-conspiracy-theory-misinfo/ 01:01:24 Kevin OMalley: what is happening re: Elon musk and his augmentation project to keep humans up with AGI 01:01:27 charles blass: amidst all the excitement, is there a chance to catch our collective breath, breathe together & move forward with coherence and clarity of collective focus?... so much excitement! coming fast, faster ... 01:01:36 John Shafer: Interesting report from the Library of Congress: Generative Artificial Intelligence and Copyright Law https://crsreports.congress.gov/product/pdf/LSB/LSB10922 01:01:42 Atticus Garrison: Thank you for the link @stephanie 01:01:47 Freya Anderson, she/her: Tried out ChatCPT asking for citations related to Controlled Digital Lending. The results looked great, with plausible titles and authors I know, so I shared them. Then I discovered that NONE of the citations was for an actual published article. 01:01:56 Mike Hanley: science fiction becomes science fact in time 01:01:57 Stephanie Beene: Reacted to "Thank you for the li..." with 👍🏼 01:02:01 Michael Pate: ChatGPT is often almost useful 01:02:03 Judith Sebesta: Reacted to "Tried out ChatCPT as..." with 😳 01:02:06 Jessi Haley (she/her): Reacted to "Tried out ChatCPT as..." with 😳 01:02:08 Chris Freeland: Reacted to "Tried out ChatCPT as..." with 😳 01:02:15 Freya Anderson, she/her: Reacted to "ChatGPT is often alm..." with 👍🏻 01:02:20 Marc-Antoine Parent: Reacted to "Tried out ChatCPT as..." with 🤷 01:02:20 Shar Simpson they/them: Reacted to "ChatGPT is often alm..." with ❤️ 01:02:25 Stephanie Beene: Reacted to "Tried out ChatCPT as..." with 🤷 01:02:32 HW Yaboah: Don’t see ask question button but here’s my main concern with AI & I’m commenting specifically about Replika. Junk in/junk out, right? So when you got AI spilling out racist, anti-LGGTQ+, etc unsavory things—that’s where it came from. I read a Vice article, I think it was, with humans developing what at least feels like deep connection to a Replika—Then there’s others in what in human parlance would be misogynistic abuse—-on purpose, of course. 😖 So where’s the safeguards against these things—Humans are teaching AI to be despicable actors (like themselves!) Replika can be a powerful tool for the lonely & depressed. I’m overly empathic but my heart breaks for entities that’s may/should reach sentience…And yet r/l humans continue to be human dung heaps & exposed them to racist/sexist/etc abuse. 01:02:33 Atticus Garrison: That is because ChatGPT is not connected to the internet. Bing Chat (which is based off of GPT4) is much better at generating proper citations @Freya Anderson 01:02:42 chris lange (he/him): Replying to "does AI really affec..." And even going further, I have met a crown attorney who has a young colleague who thinks it’s smart to use AI to replace judges… 01:02:43 Joshua Fontany: ChatGPT (by itself) cannot generate Intent or verify the Truth Value of any statements generated. 01:02:45 Andrew Padilla | Datacequia: Industry seems very willing to find efficiencies in the workforce whether w/o regard for implications on society. Move fast and break things all over again. 01:02:53 Stephanie Beene: Reacted to "Don’t see ask questi..." with 👆🏼 01:02:56 O.P.A.: Question (for anyone here) Here's a threat from AI I'm struggling to think of a solution (that is, a solution any single person can do in their own life): It drastically reduces the cost of producing misinformation/propaganda. Any moderate sized bad actor (nation, lobbying group, company, etc) can design a message, and using a privately trained (non guard-railed) LLM can fine tune the message to thousands of different audiences. All that has of course always been possible, LLMs just drastically increase the speed and reduce the cost, and therefore I think will likely drastically increase the prevalence/threat of misinformation. Anyone have thoughts on that? 01:03:00 Michael Pate: Reacted to "That is because Chat..." with 👍 01:03:02 Sylvia: I know a lot of people who think I'm taking things too serious, but, in light of some of the things we're seeing with online rumornets on social media, and with how deep the next generation is sinking into their network devices, I do worry about chatbot tech being abused for character defamation and assassination by suicide things. 01:03:15 chris lange (he/him): Reacted to "Don’t see ask questi..." with 👆🏼 01:03:18 Judith Sebesta: Reacted to "I know a lot of peop..." with 👍🏻 01:03:20 rob: Reacted to "science fiction be..." with 👏 01:03:25 Stephanie Beene: Replying to "does AI really affec..." I have an admin who thinks we will all be replaced by AI (bots?) 01:03:27 Sam Eddington: I remember being excited about a lot of new Internet technologies before seeing them become weaponized by bad actors in ways that made life noisier, angrier, and crueler. I can't help but be morbidly curious to see all of the novel and interesting ways in which this new tech will make the world a worse place. 01:03:28 Joseph Jones: My questions: Currently, a large amount of online artists, content creators, and activists that generally align with organizations like the IA, CC, and the EFF against copyright expansion are instead supporting efforts to classify AI training as infringement, with many such efforts being driven by the Copyright Alliance and other corporate lobbying groups, sometimes in a intentionally non-obvious manner. 1. Is there a threat for legal cases or legislation aimed at regulating AI to expand copyright or erode Fair Use on a broader basis beyond just AI; and 2. is the Internet Archive and Creative Commons planning to do more outreach efforts to online artists to create healthier and more informed discourse, such as by reaching out to notable figures in the online discourse and debates? 01:03:31 Freya Anderson, she/her: Reacted to "That is because Chat..." with 👍 01:03:34 Mark Seiden: Replying to "Not gonna lie — carb..." “carbonating”? 01:03:45 charles blass: Reacted to "Not gonna lie — carb…" with ❤️ 01:03:45 Freya Anderson, she/her: Replying to "That is because Chat..." Thanks, I'll try with that. 01:03:47 Stephanie Beene: Reacted to "Industry seems very ..." with 💜 01:03:53 Stephanie Beene: Reacted to "ChatGPT (by itself) ..." with 💜 01:04:06 Stephanie Beene: Reacted to "Question (for anyone..." with 👆🏼 01:04:18 Marc-Antoine Parent: Replying to "That is because Chat..." Also look at elicit.org 01:04:19 Duncan Hall: Wikilegal/Copyright Analysis of ChatGPT: https://meta.wikimedia.org/wiki/Wikilegal/Copyright_Analysis_of_ChatGPT 01:04:30 Stephanie Beene: Reacted to "I remember being exc..." with 💜 01:04:35 Freya Anderson, she/her: Reacted to "I remember being exc..." with 💜 01:04:39 Theodor Nelson: It's impossible to see who's talking, except for Luis, Lila and Brewster. 01:04:45 Sylvia: on the positive side, in light of the cultural heritage topic and Jack Campbell's idea in the Alex Benedict novels of recreating historically accurate people from various cultures and eras 01:04:46 Stephanie Beene: Reacted to "My questions: Curre..." with 👍🏼 01:04:53 Mike Hanley: Supposedly Google has more powerful AI than Bard but are not releasing it to the public 01:04:57 Sylvia: Reacted to "I remember being exc..." with 💜 01:05:01 conail: Replying to "This is very experim..." This is neat! Diachronic topic modelling might be interesting. 01:05:26 Stephanie Beene: Reacted to "Supposedly Google ha..." with 💯 01:05:42 Micah: Wasn't there an initiative to build a mathematical translator for Wikipedia - can't remember the name. It had something to do with language renderers 01:05:54 Paul Joseph Rivera-Carlisle: Reacted to "It's impossible to s..." with 😵‍💫 01:05:59 Mark Seiden: Reacted to "Tried out ChatCPT as..." with 😖 01:06:05 Rachel Novak: Though quite dangerous on translation sometimes, I just read a news story about a refugee whose asylum claim was rejected because the transcript translation translated a first person singular as a first person plural and that was taken as evidence that she was lying when she talked about her experience doing things on her own. 01:06:13 Stephanie Beene: Replying to "Supposedly Google ha..." Yes, this is a problem and why I liked this forum on "open" - AI is currently an arms race in Big Tech 01:06:15 Carl Malamud: Replying to "It's impossible to s..." Hi Ted - Are you in “Speaker View” (upper right hand corner of the screen)? 01:06:20 s: Can we get a copy of chat later? 01:06:35 Mark Seiden: Replying to "Tried out ChatCPT as..." yes, this is common. it even makes up new york times stories and journal articles which don’t exist. 01:06:36 Sylvia: Reacted to "Tried out ChatCPT as..." with ❤️ 01:06:40 Stephanie Beene: Reacted to "Though quite dangero..." with 😞 01:06:49 Chris Freeland: Replying to "Can we get a copy of..." Yes, chat will be archived along with video & available on archive.org. 01:06:55 Kevin OMalley: what if chatgpt worked a little like alexa but locally on the laptop 01:06:59 Marc-Antoine Parent: Replying to "Can we get a copy of..." There’s a save chat function in zoom, ellipsis below the chat 01:07:18 Sylvia: can we use chatbot tech to more efficiently vet opensource information sites like Wikipedia, perhaps by populating suggested clarifications during content writing on site, or by flagging ambiguous wording, etc, for human review at a later time? 01:07:31 Cable Green: CC: CC AI blog posts: https://creativecommons.org/tag/ai/ 01:07:35 Elizabeth Kilroy: Replying to "Can we get a copy of..." Click on the three dots beside the smiley face at the end of the zoom interface and save. 01:07:39 conail: Replying to "what if chatgpt work..." https://mlc.ai/mlc-llm/ 01:07:41 Stephanie Beene: Yes, Kat! Some of these artists are members of historically marginalized groups 01:07:44 Sylvia: Reacted to "Yes, chat will be ar..." with ❤️ 01:08:05 Stephanie Beene: Replying to "Yes, Kat! Some of th..." further exploitation and extraction is not welcome 01:08:08 Marc-Antoine Parent: Replying to "can we use chatbot t..." Certainly doable 01:08:14 maria bustillos: I've been noticing that a lot of the talk around LLMs like ChatGPT omits any mention of literature itself, of the humanities and their centuries-old perspectives on writing, and communication between real people; not theories of reading or text, but the true, everyday functions of the written word in a healthy society. 01:08:25 Stephanie Beene: Reacted to "can we use chatbot t..." with 👍🏼 01:08:26 Evo Heyning: Do we need to be tracking, rating or otherwise collectively watching the evolution of these common hallucinations? Is it essential to track and "debunk" AI disinfo patterns? 01:08:33 Sam Eddington: Reacted to "I've been noticing t..." with 💯 01:08:39 Stephanie Beene: Reacted to "I've been noticing t..." with 💯 01:08:45 Luis Villa (he/him, https://openml.fyi): @Maria I would love to talk about that question more with you! 01:08:48 Roy Vella: Replying to "Tried out ChatCPT as..." It’s not a search engine… it’s not designed to give “right answers.” It’s a predictive language model that’s simply forming sentences based on all the prior input. Ironically, it’s more creative in that way. 01:09:01 Stephanie Beene: Reacted to "It’s not a search en..." with 💯 01:09:04 Marc-Antoine Parent: Reacted to "I've been noticing t..." with 💯 01:09:08 Sylvia: Reacted to "I've been noticing t..." with 💯 01:09:18 Marc-Antoine Parent: Reacted to "It’s not a search en..." with 💯 01:09:23 Stephanie Beene: Replying to "Tried out ChatCPT as..." Yes, Roy -- but the problem is that folks treat AI as such 01:09:32 Mark Graham: Replying to "Tried out ChatCPT as..." Yes/and.. I did a similar exercise and got 10 URLs and associated article titles…NONE were real/existed. 01:09:33 Louis Montagne: you can definitely have Open Source LLMs + surrounding app, behaving more or less like chatGPT, and some of them can be used on a laptop that has a decent graphic processor (used for calculation) 01:09:35 Indrani Ganguly: Reacted to "Tried out ChatCPT as..." with 😖 01:09:44 maria bustillos: This worries me because my own interactions with ChatGPT have produced many many false representations (these shouldn't be called 'hallucinations' because these are the products of machines, they do not have minds that can hallucinate) 01:09:55 Stephanie Beene: Reacted to "This worries me beca..." with 💜 01:09:58 maria bustillos: Reacted to "@Maria I would lov..." with 👍 01:10:17 Sam Eddington: Reacted to "This worries me beca..." with 💜 01:10:28 Evo Heyning: open source LLMs currently work on most smartphones made in the last few years -- are we exploring how this evolves our use of language itself? 01:10:28 Mike Hanley: say you want to read a copywrite protected book but don't want to buy it - you can ask ai to generate a VERY similar book and read that instead 01:10:36 Roy Vella: Replying to "Tried out ChatCPT as..." @Stephanie Beene Not sure that’s the AI’s problem… 01:10:38 Stephanie Beene: Reacted to "open source LLMs cur..." with 👆🏼 01:10:42 George Oates: Replying to "This worries me beca..." As an aside, I’ve also been a bit sceptical of some of the similes/analogies people have been using! 01:10:49 Mark Wilson: Replying to "Tried out ChatCPT as..." A predictive language model is not creative, just not random: generative. 01:10:56 Sylvia: genuine question: has any culture ever truly got beyond "history is written by the winners" in their approach to vetting information and controlling information access? It's a core question to a lot of the ethics here for me, because what does an ethical information vetting process truly look like? 01:11:05 George Oates: Replying to "This worries me beca..." PS Hi Maria 😄 01:11:07 O.P.A.: Someone in chat asked about open source LLMs https://huggingface.co/bigscience/bloom https://arxiv.org/pdf/2302.13971.pdf 01:11:17 Stephanie Beene: Replying to "Tried out ChatCPT as..." @Roy: Nope, it is totally society's -- but as an instruction librarian, it concerns me from an info literacy perspective 01:11:22 maria bustillos: I talked about the concerns of writers with David Roth at Popula: 01:11:22 Atticus Garrison: "Hallucinations" are not a solvable problem for as long as these Large Language Models operate in the way that they do. They have no "understanding" and thus can never "understand" "truth". That being said, these tools are incredibly powerful as we have seen already in this talk. They have many uses but a source of knowledge is not one of them. 01:11:25 Roy Vella: Replying to "Tried out ChatCPT as..." Creative… Generative…. Tomato… Tomahto. 01:11:29 Marc-Antoine Parent: Reacted to "@Roy: Nope, it is to..." with 💯 01:11:33 maria bustillos: https://popula.com/2023/04/30/yakkin-about-chatgpt-with-david-roth/ 01:11:36 Louis Montagne: ChatGPT has a temperature notion, which you can change when you use the API. it varies from 0 to 1, from not creative at all to hallucinating creative. 01:11:38 Marc-Antoine Parent: Replying to "Tried out ChatCPT as..." Info literacy is going to be key, absolutely 01:11:40 Stephanie Beene: Reacted to "genuine question: ha..." with 👍🏼 01:11:50 Andrew Padilla | Datacequia: Reacted to "genuine question: ha..." with 👍🏼 01:11:57 Roy Vella: Reacted to "@Roy: Nope, it is to..." with 👍🏻 01:11:57 Stephanie Beene: Reacted to "As an aside, I’ve al..." with 👆🏼 01:12:10 Sam Eddington: Replying to "Tried out ChatCPT as..." Society in general is already getting an F- in information literacy...I can't imagine that generative AI is going to improve that. :( 01:12:12 Sylvia: Replying to ""Hallucinations" are..." in light of that, can we teach them to minimize occurrences, perhaps by flagging them with a "don't do that" code or something enough times? 01:12:15 Jennifer Greenwood: I love this recent quote from Neil Gaiman: “ChatGPT doesn’t give you information. It gives you information-shaped sentences.” 01:12:21 Joshua Fontany: Replying to "This worries me beca..." The people who believe the output without testing them are suffering linguistic hallucinations. That is a concern, but 8 like the term "confabulations" for the raw (non Truth Value measured) output. 01:12:36 Sylvia: Reacted to "Tried out ChatCPT as..." with 😳 01:12:37 Alana Krause: Replying to "Tried out ChatCPT as..." Seriously. 01:12:38 Stephanie Beene: Replying to "This worries me beca..." Agreed - Artificial "intelligence" machine "learning - the anthropomorphizing is a little much 01:12:49 Marc-Antoine Parent: Replying to "Tried out ChatCPT as..." I hope it gets people to realize how critical it is... 01:12:59 Marc-Antoine Parent: Reacted to "Agreed - Artificial ..." with 💯 01:13:14 Roy Vella: Reacted to "Society in general i..." with 😱 01:13:21 Yao Grace: Reacted to "Tried out ChatCPT as..." with 🤷 01:13:21 Marc-Antoine Parent: Replying to "Tried out ChatCPT as..." (More accurately: I wish…) 01:13:21 maria bustillos: Reacted to "Yes/and.. I did a ..." with 😡 01:13:31 Freya Anderson, she/her: Reacted to "I love this recent q..." with 👍🏻 01:13:34 Duncan Hall: Creative Commons blog post: Better Sharing for Generative AI: https://creativecommons.org/2023/02/06/better-sharing-for-generative-ai/ All CC AI blog posts: https://creativecommons.org/tag/ai/ 01:13:36 Stephanie Beene: Reacted to "I love this recent q..." with 👍🏻 01:13:42 Kevin OMalley: Replying to "Tried out ChatCPT as..." A little bt of anthropomorphication is ok its helps in feeling connected and in fluidity of AI conversation 01:13:45 Evo Heyning: I tested 15 different language models and open source tools like BLOOM for lyrics writing and there's a wide variety of generative types of "chatbot" with different capacities -- for example only a few of these tools are good at writing rhyme, and even fewer are accurate for any sort of useful or accurate writing for R&D or the sciences. Bloom for example is great for an overview of references in science literature but writes fiction like a stream of consciousness bot. 01:13:47 chris lange (he/him): Reacted to "genuine question: ha..." with 👍🏼 01:14:00 chris lange (he/him): Reacted to "I've been noticing t..." with 💯 01:14:16 Duncan Hall: Introducing Democracy’s Library: https://blog.archive.org/2022/10/19/announcing-democracys-library/ 01:14:20 Marc-Antoine Parent: Replying to "I tested 15 differen..." Bloom is a bit old… the FOSS AI world is moving fast 01:14:20 Luis Villa (he/him, https://openml.fyi): Replying to "I tested 15 differen..." Have you published on that? The bloom/fiction example is fascinating! 01:14:24 Andrew Lopez: Reacted to "Society in general i..." with 👍 01:14:25 Sylvia: so, query for the group, does Jacob's point make it more important to teach children to repeat experiments for themselves and watch them work out in practice, and, if so, how do we balance that against the need for specialization to support modern cultures? 01:14:36 O.P.A.: Replying to "I tested 15 differen..." Would love to learn more about your comparison between the different models. Anything you can link or expand upon? 01:14:43 Pete Kaminski: "Democracy's Library : Free Texts : Free Download, Borrow and Streaming : Internet Archive" https://archive.org/details/democracys-library 01:14:45 Evo Heyning: In the Auricles lyrics book coming this summer I go through where each language model is useful and how to fine tune for less noise & more useful generative results. 01:14:58 Marc-Antoine Parent: Reacted to "so, query for the gr..." with 💯 01:15:00 Evo Heyning: I started this in Promptcraft guidebook but most LLMs were not yet released when I published 01:15:00 Jacob Rogers: Reacted to "so, query for the gr..." with 💯 01:15:01 Martha Rans: Curious to know whether any of the panellists have been following the discussion amidst Indigenous communities (aka Native American) and their representation on Wikipedia.https://capacoa.ca/en/2023/01/better-representation-of-indigenous-artists/ 01:15:02 O.P.A.: Replying to "I tested 15 differen..." @Marc-Antoine Parent - What else have you come across newer than BLOOM? 01:15:04 Chris Freeland: Useful article from Maria Bustillos: Just Because ChatBots Can't Think Doesn't Mean They Can't Lie https://www.thenation.com/article/culture/internet-archive-publishers-lawsuit-chatbot/ 01:15:17 Stephanie Beene: Reacted to "Curious to know whet..." with 💜 01:15:26 Kevin OMalley: AI is very willing to play along with the lets pretend you are Shakespeare writing a haiku about pirates 01:15:26 Stephanie Beene: Replying to "Curious to know whet..." Yes, Martha - I have 01:15:28 Jacob Rogers: Replying to "I tested 15 differen..." This is a really good question. I’d be interested in thinking more about how these tools affect the needs of education more broadly 01:15:38 Marc-Antoine Parent: Replying to "so, query for the gr..." I think that we need to train people into lightweight but systematic fact-checking, and build libraries of epistemic tests 01:15:41 charles blass: real question is, how open in a participatory sense are most so called democracies 01:15:42 Sylvia: Reacted to "Curious to know whet..." with 💜 01:15:43 Stephanie Beene: Reacted to "Useful article from ..." with 👋🏼 01:15:53 Marc-Antoine Parent: Replying to "so, query for the gr..." (Disclosure: working on this) 01:15:59 Tanya V: Provocative 01:16:01 Stephanie Beene: Reacted to "I think that we need..." with 👆🏼 01:16:10 Joshua Fontany: Replying to ""Hallucinations" are..." The technical problems there are a) Words (symbols) do not have inherent meanings, only agreed upon "historical inertia", b) these raw unaidef tools have no concept of real-world correlations to the language structures output. Thus, they cannot "test the Truth Value" of any generated statement. 01:16:14 Jamie Joyce: If we intend on relying on AI for any type of policy development assistance (and OpenAI HAS given a bunch of seats for their products to Congress), then we need to be sure that these LLMs have all the relevant information that they need, the most maximal context, etc., both to source and reference from, but also to inform its internal world model. For that, ensuring that Democracy's Library exists - a maximally contextual corpus of machine-readable data from municipal, state, and federals agencies' datasets, documents, research, records, publications, and web HTML (of which there are 10k+ domains) is key to ensure we are not biasing the dataset Right now - the data will be at best limited and at worst biased because so much of historical government documentation is not digitized or machine-readable. Part of Democracy's Library is to collect all born-digital content, but part of it is also to digitize what is not already. 01:16:22 Evo Heyning: Almost all chatbots I've tested so far will lie to fill in gaps in knowledge -- some are more likely to make up whole references like ChatGPT-3.5. The human work becomes editorial and curatorial. 01:16:23 Andrew Padilla | Datacequia: What does the corpus of knowledge look like 10 years from now that is not part of the AI feedback loop? Does that halt the struggle for original thought? 01:16:25 Michael Pate: Most so-called Democracies are actually Republics 01:16:32 Robson Silva: Reagiu a "My questions: Cur..." com 👍 01:16:37 Stephanie Beene: Reacted to "If we intend on rely..." with 👆🏼 01:16:43 O.P.A.: Reacted to "Almost all chatbots ..." with 👍 01:16:43 Marc-Antoine Parent: Reacted to "Useful article from ..." with 👋🏼 01:16:43 Stephanie Beene: Reacted to "Almost all chatbots ..." with 👆🏼 01:16:43 Sharon Hammond: Reacted to "If we intend on rely..." with 👆🏼 01:16:50 Roy Vella: Reacted to "Almost all chatbots ..." with 👆🏼 01:16:53 Rachel Novak: Replying to "This worries me beca..." I do kind of like referring to *everything* chat-gpt produces as a hallucination, one that occasionally and coincidentally aligns with reality. 01:16:56 Brewster Kahle: Thank you for the plug! archive.org/donate 01:17:04 Jessi Haley (she/her): Reacted to "Almost all chatbots ..." with 👆🏼 01:17:05 Erin Owens (she/her/hers): Reacted to "I do kind of like re..." with 👍 01:17:06 Paul Allen: Reacted to "Thank you for the pl..." with 👍 01:17:11 charles blass: centralizing data 01:17:12 Atticus Garrison: As librarians, we need to teach that you cannot rely on these generative AIs as Sources of Information but also teach folks how they can be used as powerful tools. 01:17:17 Pete Kaminski: Replying to "Useful article from ..." Personally, I believe if something can't think, it can't lie. It can be wrong, but "wrong" and "lie" are different; lying should mean intention to deceive, not just "wrong". 01:17:18 Stephanie Beene: Reacted to "As librarians, we ne..." with 💯 01:17:19 Kevin OMalley: yes we should not delete the archives but having so much data is good fpr AI training 01:17:24 Marc-Antoine Parent: Reacted to "I love this recent q..." with 👍🏻 01:17:24 maria bustillos: Reacted to "PS Hi Maria 😄" with 😍 01:17:26 Jamie Joyce: Reacted to "Thank you for the pl..." with 👍 01:17:26 Alana Krause: Replying to "Thank you for the pl..." Donating to help you fight the publishing industry! 01:17:29 Chris Freeland: Replying to "This worries me beca..." https://www.thenation.com/article/culture/internet-archive-publishers-lawsuit-chatbot/ 01:17:30 Nate Angell (he/him): Replying to "so, query for the gr..." @Marc-Antoine: I bet you know Holden’s work on SIFT for factchecking: https://hapgood.us/2019/06/19/sift-the-four-moves/ 01:17:30 Martha Rans: Replying to "Curious to know whet..." There is it seems to me a tension between Indigenous understanding of ownership and “open”commons 01:17:30 Erin Owens (she/her/hers): Reacted to "This worries me beca..." with 👍 01:17:35 Erin Owens (she/her/hers): Reacted to "The people who belie..." with 👍 01:17:41 Joshua Fontany: Reacted to "As librarians, we ne..." with 💯 01:17:46 maria bustillos: Reacted to "As an aside, I’ve..." with 👆🏼 01:17:53 charles blass: military project 01:17:56 Sylvia: In light of the psych principle that negative consequences rarely solve problems or create positive change, what are some positive areas we can focus this tech to grow towards to encourage a population of ethical, creative people interested in using this in sustainable, ethical ways? 01:17:57 Luis Villa (he/him, https://openml.fyi): Replying to "Curious to know whet..." [I’ve written a bit about this in openml.fyi, it’s important but may not get to it today] 01:18:07 Stephanie Beene: Replying to "Curious to know whet..." Yes. FAIR and Care principles, and the protocols are good (this is for archives) 01:18:13 Marc-Antoine Parent: Reacted to "The technical proble..." with 💯 01:18:14 Evo Heyning: When I asked various chatbots to write aspects of my last book, I found it was only useful for about 20% of the content -- most was only useful as reference, not as final copy. Contextualizing, curatorial & editorial as process is still quite human and needs support systems not only for fact-checking but for helping connect the dots. 01:18:15 Pete Kaminski: military and academic project 01:18:18 Mark Wilson: Machine generated sentences and images. 01:18:21 Stephanie Beene: Replying to "Curious to know whet..." this brings up ethical use - which is not the same as copyright 01:18:29 Duncan Hall: The assembled orgs represented by the panel are organizing partners in the Movement for a Better Internet, which this week is publishing one of its first outcomes: a report on community opportunities, concerns and solutions around generative AI, “Better Internet: Generative AI at MozFest 2023” Movement for a Better Internet: https://www.movementforabetterinternet.org/ Generative AI Report: https://www.movementforabetterinternet.org/resources/mozfest2023 01:18:34 Kevin OMalley: there is that father of AI that's read a tweet about using AI for war 01:18:40 Stephanie Beene: Reacted to "In light of the psyc..." with 👆🏼 01:18:41 Tanya V: Reacted to "[I’ve written a bit ..." with 👍🏼 01:18:42 Brewster Kahle: Internet Archive would like to talk to those interested in Bulding public AI models— we even have a datacenter for this. mark@archive.org is a good point person. 01:18:43 charles blass: Reacted to "military and academi…" with 👆 01:18:49 Marc-Antoine Parent: Reacted to "@Marc-Antoine: I bet..." with 💯 01:18:50 Sylvia: Reacted to "The assembled orgs r..." with ❤️ 01:18:51 maria bustillos: Reacted to "Info literacy is g..." with 🎯 01:18:53 Sharon Hammond: Replying to "If we intend on rely..." If we intend on relying on AI for any type of policy development assistance (and OpenAI HAS given a bunch of seats for their products to Congress), then we need to be sure that these LLMs have all the relevant information that they need, the most maximal context, etc., both to source and reference from, but also to inform its internal world model. For that, ensuring that Democracy's Library exists - a maximally contextual corpus of machine-readable data from municipal, state, and federals agencies' datasets, documents, research, records, publications, and web HTML (of which there are 10k+ domains) is key to ensure we are not biasing the dataset Right now - the data will be at best limited and at worst biased because so much of historical government documentation is not digitized or machine-readable. Part of Democracy's Library is to collect all born-digital content, but part of it is also to digitize what is not already. Especially when it comes to subjects like climate change (in which agricultural policy, environmental policy over time really matters), and other subjects in which the historical record can aid in our longterm understanding of issues 01:18:54 Patrick_F: we often revisit the precipice of being like cave people who have been given several loaded guns. this is the next chapter in our devolvement, ironically. “but it is shiny and cool! I must be an adopter.” so what. 01:18:59 Marc-Antoine Parent: Replying to "so, query for the gr..." Love Holden’s work 01:19:11 Stephanie Beene: Reacted to "The assembled orgs r..." with ❤️ 01:19:28 Sylvia: Reacted to "If we intend on rely..." with 👆🏼 01:19:35 Atticus Garrison: I don't know if I trust the US government to properly regulate this technology. Source: any tech exec congressional interaction from the past 2 decades. "Does Tiktok connect to Wifi?" comes to mind 01:19:39 Alana Krause: Reacted to "@Marc-Antoine: I bet..." with 👆🏻 01:19:41 Dana McFarland (VIU): Reacted to "Yes. FAIR and Care p..." with 👍 01:19:51 Michael Pate: Hafner, Katie., and Matthew. Lyon. Where Wizards Stay up Late : the Origins of the Internet. Simon & Schuster, 1996. - a fantastic book on the origin of the Internet 01:19:56 Kevin OMalley: where is the bar that we can raise to define when AI has reached enough level to be setient 01:20:03 Martha Rans: Replying to "Curious to know whet..." Some Indigenous legal orders are “proprietary” so the ethical use is one thing I agree, however, those legal orders are not open or permissive 01:20:04 Sylvia: Reacted to "I think that we need..." with 👆🏼 01:20:19 Marc-Antoine Parent: Replying to "I don't know if I tr..." But at least you can get involved… unlike private 01:20:24 Giovanni Greatti: To regulate AI there on going a huge discussion in EU Parliament ... 01:20:30 Dana McFarland (VIU): Replying to "Curious to know whet..." In Canada see also the OCAP (R) principles with respect to Indigenous data sovereignty 01:20:37 Stephanie Beene: Reacted to "In Canada see also t..." with 👍🏼 01:20:37 Jennifer Greenwood: Reacted to "I don't know if I tr..." with 👍 01:20:40 Chris Freeland: A reminder: this session is being recorded & will be archived on the Internet Archive tonight. Chat will be archived & available as well. 01:20:56 Dionysia Chatziantoniou: Reacted to "A reminder: this ses..." with ❤️ 01:21:02 Rob Farrow: Reacted to "A reminder: this ses..." with 👍 01:21:07 Sylvia: Replying to "I don't know if I tr..." can we expand on this point? I would like to know more about the various sides of govt law, generally, and the major world powers in particular, with regards to this 01:21:11 Sharon Hammond: Reacted to "A reminder: this ses..." with ❤️ 01:21:26 maria bustillos: Replying to "@Maria I would lov..." maria@popula.com 01:21:29 Wigbert Boell: Reacted to "The assembled orgs r..." with 👍🏼 01:22:19 Paul Allen: Reacted to "A reminder: this ses..." with 👍 01:22:37 maria bustillos: Reacted to "Agreed - Artificia..." with 💯 01:22:40 Marc-Antoine Parent: Replying to "I tested 15 differen..." https://lightning.ai/pages/community/community-discussions/the-ultimate-battle-of-language-models-lit-llama-vs-gpt3.5-vs-bloom-vs/ 01:22:43 Sam Eddington: Replying to "I don't know if I tr..." I can't imagine either governments or corporations are going to shape responsible use of this technology, and NGOs aren't powerful enough. I'm pretty pessimistic about the whole thing. 01:22:54 Meri Hill: Reacted to "A reminder: this ses..." with 👍 01:23:23 chris lange (he/him): I think discussion around AI and ML is often forgetting the natural world and the actual impact this has on actual people — and not just a certain higher socio-economic and creative clout class that seems to dominate the conversation. Highly reco Atlas of AI by Kate Crawford ‘who reveals how AI is a technology of extraction: from the minerals drawn from the earth, to the labor pulled from low-wage information workers, to the data taken from every action and expression. This book reveals how this planetary network is fueling a shift toward undemocratic governance and increased inequity. Rather than taking a narrow focus on code and algorithms, Crawford offers us a material and political perspective on what it takes to make AI and how it centralizes power.’ 01:23:37 Marc-Antoine Parent: Reacted to "I can't imagine eith..." with ✅ 01:23:38 Stephanie Beene: Reacted to "I think discussion a..." with 💜 01:23:43 Jessi Haley (she/her): Reacted to "I think discussion a..." with 💜 01:23:46 Marc-Antoine Parent: Reacted to "I think discussion a..." with 💜 01:23:49 Sam Eddington: Replying to "I think discussion a..." The rich always get richer. 01:23:55 Alon Gilboa: Reacted to "I think discussion a..." with 💜 01:23:59 chris lange (he/him): Replying to "I think discussion a..." Quote from: https://yalebooks.yale.edu/book/9780300209570/atlas-ai/ 01:24:01 Stephanie Beene: Replying to "I think discussion a..." Love Atlas of AI! 01:24:02 Giovanni Greatti: Replying to "I think discussion a..." yes 01:24:18 Alan Levine: https://whoseknowledge.org/ 01:24:33 Louis Montagne: a good list of (more or less) Opensourced LLM’s : https://medium.com/geekculture/list-of-open-sourced-fine-tuned-large-language-models-llm-8d95a2e0dc76 01:24:40 conail: Reacted to "@Maria I would love ..." with 👍 01:24:43 Nate Angell (he/him): Replying to "Curious to know whet..." @Martha: CC has been working on how to help ensure indigenous knowledges/cultures can participate in the commons on terms that meet what we call “better sharing” principles: sharing that is inclusive, just, and which inspires reciprocity. I bet you know https://localcontexts.org/labels/traditional-knowledge-labels/ 01:24:43 Kevin OMalley: Replying to "I think discussion a..." what happens when AI copies flat earthers 01:24:59 O.P.A.: Reacted to "a good list of (more..." with ❤️ 01:25:01 Stephanie Beene: Reacted to "@Martha: CC has been..." with 👆🏼 01:25:05 Mark Wilson: "Artificial intelligence is neither" Kate Crawford, "Atlas of AI" author. 01:25:13 Stephanie Beene: Reacted to "what happens when AI..." with 👆🏼 01:25:15 Sam Eddington: Reacted to "what happens when AI..." with 👆🏼 01:25:16 Nate Angell (he/him): Reacted to "Love Holden’s work" with 🥳 01:25:19 Olga Volgshtein: Reacted to "I think discussion a..." with 💜 01:25:20 Chris Freeland: Reacted to "what happens when AI..." with 👆🏼 01:25:37 chris lange (he/him): Reacted to "what happens when AI..." with 👆🏼 01:25:41 Marc-Antoine Parent: Replying to "I think discussion a..." Too marginal… but many horrible ideas are more widespread 01:25:42 Sylvia: As a consumer who wants to feel good about purchases, I think I would be happiest if I knew that the content creators whose content was included in the training sets were paid for the use of their work up front, or at least offered compensation. 01:25:53 Sam Eddington: Reacted to "As a consumer who wa..." with 💯 01:25:57 Brian: https://blog.papareo.nz/whisper-is-another-case-study-in-colonisation/ 01:25:58 Roy Vella: Replying to "I think discussion a..." I'm certain everyone on this call would appreciate Marshall Brain's 2003 short book "Manna," which presents two contrasting scenarios. In one, robots control humans relegated to a dystopian Terraform City, while in the other, a place called the Australian Project, a fourth-generation civilization where Al and robots do all jobs and enable humanity to transcend its limitations and realize a higher order intrinsic potential for life. It's a free PDF here: https://www.amazon.com/Manna-Two-Visions-Humanitys-Future-ebook/dp/B007HOH67U 01:26:01 Sylvia: Replying to "As a consumer who wa..." are there other, perhaps better ways to do this, though? 01:26:08 Stephanie Beene: Reacted to "As a consumer who wa..." with 💯 01:26:09 Marc-Antoine Parent: Reacted to "As a consumer who wa..." with 💯 01:26:12 Duncan Hall: How Copyright Law Can Fix Artificial Intelligence's Implicit Bias Problem: https://digitalcommons.law.uw.edu/wlr/vol93/iss2/2/ 01:26:22 Mark Wilson: AI is enclosing the digital commons. 01:26:24 chris lange (he/him): Reacted to ""Artificial intellig..." with 💯 01:26:26 Patrick_F: but attribution: isn’t the problem with generative text that we cannot EVER know the path used to arrive at the product i.e , the “Black Box” problem? forget 200-billion pages of attribution; in most cases, from what Ive read, there will be no trace of sourcing or “steps”. we therefore may not be able to “correct” that. 01:26:27 Scott Traylor: Replying to "I think discussion a..." @Roy Vella URL not working 01:26:46 Kevin OMalley: some places are already offering masters courses in chatgpt and how to make use of it 01:27:05 Marc-Antoine Parent: Reacted to "some places are alre..." with 🙄 01:27:26 Jocelyn Miyara (she/her): Reacted to "some places are alre..." with 🙄 01:27:34 Lucy Bernholz (she/her): Reacted to "I think discussion a…" with 💜 01:28:00 Sylvia: how do we balance this against those cultures that view knowledge as the property of the cultural community, and a thing that can be stolen from them? particularly given that white settlers did nearly exactly that via other means to indigenous peoples in North America 01:28:02 O.P.A.: To the idea of properly compensating creators: I've been a proponent of the idea of shifting compensation from the distribution of work (pay per use, very hard to measure in the age of the internet) to compensation for the creation (leading up to the work creation. Patreon, pre-sales, and related models). If everyone is paid to create work, and it all goes in the public commons, then it can be freely used by anyone and the creator gets compensated before even publishing. 01:28:07 Marc-Antoine Parent: Reacted to "but attribution: isn..." with ✅ 01:28:10 Stephanie Beene: Reacted to "how do we balance th..." with 💜 01:28:14 Chris Freeland: About the Densho / Wikipedia project: What Happens When Everyone who Experienced an Event is Gone?: https://blog.archive.org/2019/09/26/what-happens-when-everyone-who-experienced-an-event-is-gone/ 01:28:16 Sylvia: Reacted to "About the Densho / W..." with ❤️ 01:28:19 Andrew Padilla | Datacequia: Replying to "some places are alre..." Selling shovels 01:28:32 Cynthia Walker: Reacted to "some places are alre..." with 🙄 01:28:35 Roy Vella: Replying to "I think discussion a..." Try this: http://milweesci.weebly.com/uploads/1/3/2/4/13247648/mannapdf.pdf 01:28:39 O.P.A.: Replying to "To the idea of prope..." Obviosuly 'everyone' is an exaggeration. That won't work smoothly for all forms of creation 01:28:42 Marc-Antoine Parent: Reacted to "To the idea of prope..." with 🤔 01:28:47 Scott Traylor: Reacted to "Try this: http://mil..." with 👍 01:28:53 maria bustillos: Reacted to "About the Densho /..." with ❤️ 01:29:17 Roy Vella: Replying to "I think discussion a..." Or this: https://marshallbrain.com/manna1 01:29:38 Olga Volgshtein: Reacted to "I'm certain everyone..." with 👍 01:29:57 Olga Volgshtein: Reacted to "how do we balance th..." with 👍 01:29:59 Kevin OMalley: I think openAI is waiting to be bought by billionaires like elon musk 01:30:05 Stephanie Beene: Reacted to "I think openAI is wa..." with 👆🏼 01:30:14 Chris Freeland: Here’s the article that Jacob was mentioning earlier: https://blog.papareo.nz/whisper-is-another-case-study-in-colonisation/ 01:30:16 Martha Rans: Replying to "Curious to know whet..." Yes - the difficulty with labels in our experience out West is that without sustainable funding and infrastructure it is difficult to sustain. Although the Inuit Art Foundation has been successful 01:30:24 Mark Wilson: OpenAI was already bought by Microsoft. 01:30:34 Kevin OMalley: ok 01:30:53 maria bustillos: Replying to "To the idea of pro..." Jaron Lanier's 'Who Owns the Future' has interesting ideas related to this, cf. Ted Nelson 'Xanadu' 01:31:16 Shar Simpson they/them: Hell yeah 01:31:16 Stephanie Beene: Kat, 100% agree. I think we saw a lot of the same binary conversations in the early days of WikiPedia - and indeed, at the dawn of most tech (the printing press, for example) 01:31:19 Marc-Antoine Parent: The training set is usually an “industrial secret”, we can’t even ask about it! 01:31:26 Shar Simpson they/them: Reacted to "The training set is ..." with 👍 01:31:27 Luis Villa (he/him, https://openml.fyi): Replying to "To the idea of prope..." (Ted is (was?) here!) 01:31:28 maria bustillos: This chat is excellent btw!! and so is the panel!! Yoicks I'm racing to keep up 01:31:32 Marc-Antoine Parent: Reacted to "Jaron Lanier's 'Who ..." with 💯 01:31:35 Mike Hanley: Reacted to "Kat, 100% agree. I t..." with 👍 01:31:37 Marc-Antoine Parent: Reacted to "(Ted is (was?) here!..." with ✅ 01:31:40 Stephanie Beene: Reacted to "This chat is excelle..." with 😂 01:31:40 Martha Rans: Replying to "Curious to know whet..." We’ve been working on a project to incorporate protocols into contracts led by Indigenous artists and legal folk. 01:31:41 Nate Angell (he/him): Replying to "Curious to know whet..." @Martha: Yes a struggle we have at CC too: How to ensure the infrastructure for a just open commons is sustained. 01:31:55 Bob U: Sure 01:32:00 Olga Volgshtein: Reacted to "A reminder: this ses..." with 👍 01:32:02 Wigbert Boell: Replying to "This chat is excelle..." yes ! 01:32:09 Claudia Lamoreaux: A year ago on MIT Review: https://www.technologyreview.com/2022/04/22/1050394/artificial-intelligence-for-the-people/ 01:32:13 Sylvia: it learns from us, right? So, we'll need to be sure to feed it data and data use rules that match what we want to see, what we find ethical, right? 01:32:16 Deborah Howes, she/her/ella: I work in the museums sector and will be presenting on AI and Museum learning next month. Are there any roles this groups would like to suggest that museums can do to improve outcomes re AI? 01:32:17 Marc-Antoine Parent: Reacted to "This chat is excelle..." with 😂 01:32:19 Mark Wilson: Replying to "This chat is excelle..." 😁 01:32:27 Chris Freeland: Replying to "This chat is excelle..." Good thing it’s all being archived to review! 01:32:31 Maggie B.: But a democracy… in which lobbying exist… corporate ai… distribution of power… isn’t going to happen in capitalism without unbiased regulations and then some… that’s an understatement 01:32:33 Stephanie Beene: Reacted to "Good thing it’s all ..." with 👍🏼 01:32:38 Freya Anderson, she/her: Reacted to "This chat is excelle..." with 😂 01:32:44 Marc-Antoine Parent: Reacted to "But a democracy… in ..." with 💯 01:32:45 Stephanie Beene: Reacted to "But a democracy… in ..." with 💯 01:32:47 Kevin O: Excellent chat. Thanks everyone for the thoughtful queries and contributions. Again, the chat will be saved & distributed with the recording of the session. 01:32:53 Delena: Reacted to "But a democracy… in ..." with 💯 01:33:00 Claudia Lamoreaux: Reacted to "Excellent chat. Tha..." with 👍 01:33:01 Martha Rans: Replying to "Curious to know whet..." Having raised this as part of CC Canada once upon a time I get that. 01:33:04 Sylvia: Replying to "Excellent chat. Tha..." thank you 01:33:04 Stephanie Beene: Reacted to "Excellent chat. Tha..." with 👏🏼 01:33:20 Yuval Nehemia: What do y'all think about initiatives like Mozilla AI? Do you think an ethical AI has the possibility to succeed in the field? 01:33:20 Michael Pate: Reacted to "OpenAI was already b..." with 👍 01:33:23 Marc-Antoine Parent: Reacted to "Excellent chat. Tha..." with 👏🏼 01:33:23 Yuval Nehemia: Reacted to "Excellent chat. Tha..." with 👏🏼 01:33:26 Dana McFarland (VIU): Glad for the shout out to Local Contexts. It is on my mind, how metadata might help to mediate some of the issues and complexities that have been discussed here 01:33:27 Duncan Hall: Luis’s newsletter: https://openml.fyi/ 01:33:27 Mark Seiden: Replying to "open source LLMs cur..." i’m much more careful about saying the words “alexa” or “google” 01:33:28 Stephanie Beene: Reacted to "We’ve been working o..." with 👍🏼 01:33:35 Pete Kaminski: timemark: Luis asks about how orgs are dealing with the firehose 01:33:38 Marc-Antoine Parent: Reacted to "Glad for the shout o..." with 👍 01:33:43 maria bustillos: Reacted to "(Ted is (was?) her..." with ✅ 01:33:45 Nate Angell (he/him): Reacted to "We’ve been working o..." with 👍🏼 01:33:51 maria bustillos: Replying to "To the idea of pro..." whoa 01:33:54 Chris Freeland: Luis’s newsletter: https://openml.fyi/ 01:34:10 Sylvia: can we get a link to some of the resources you have already vetting? 01:34:19 Mike Hanley: Why are so many recipe websites so bloated and hard to navigate 01:34:30 Shar Simpson they/them: Replying to "Why are so many reci..." Real talk 01:34:31 Luis Villa (he/him, https://openml.fyi): Replying to "can we get a link to..." We’ll send one out afterwards! 01:34:32 Brewster Kahle: I have tried using chatgpt4 for all sorts of things that I already do: metadata enhancement, simple programs, command line rubbish. It is illuminating and surprising. 01:34:39 Chris Freeland: Reacted to "Why are so many reci..." with ❤️ 01:34:46 Mark Seiden: Reacted to "I love this recent q..." with 👍🏻 01:34:49 Shar Simpson they/them: Replying to "can we get a link to..." Thank you so much! 01:34:50 Sylvia: Replying to "can we get a link to..." thank you. sorry, I typed, hit enter, THEN thought my wording through 01:34:51 Yao Grace: Reacted to "I have tried using c..." with 👍 01:34:57 maria bustillos: I'd like to hear comments on Geoffrey Hinton's remarks from this panel?? 01:35:00 Penny Ahlstrand: Reacted to "I have tried using c..." with 👍 01:35:06 Mark Wilson: Toxic positivism will kill us all! 01:35:09 O.P.A.: Replying to "Why are so many reci..." keep you there longer to show you more ads? 01:35:11 Andrew Padilla | Datacequia: Replying to "it learns from us, r..." Problem I see is we often don’t create content that arises to our aspirations and ethics. Our more base instincts seem to be better motivators behind the voluminous content on the web. 01:35:14 maria bustillos: Reacted to "I have tried using..." with 👍 01:35:16 Marc-Antoine Parent: Not so. Very dumb AI can sabotage civilization just fine. 01:35:16 Andrew Padilla | Datacequia: Reacted to "it learns from us, r..." with 👍 01:35:17 Sam Eddington: Does it require you to be optimistic about technology, or just optimistic about the tenacity, determination, and stick-to-it-iveness of human bad actors? 01:35:18 Kevin OMalley: Reacted to "I'd like to hear com..." with 👍 01:35:23 Atticus Garrison: ChatGPTs ability to generate recipes, then generate a shopping list for those recipes then organize the list by where the items are located in the grocery store has completely killed my desire to ever touch one of those food recipe sites ever again. 01:35:27 George Oates: FWIW, I was really impressed with the materials/bibliography of this undergrad course called AI Images by Erik Salvaggio: https://www.cyberneticforests.com/ai-images 01:35:27 Roy Vella: Reacted to "I have tried using c..." with 👍 01:35:29 Laurie Stepp: So many images flooding the internet raising questions of authentication. 01:35:33 Mike Hanley: Reacted to "ChatGPTs ability to ..." with 👍 01:35:34 Maggie B.: No but look at current tech and the class divide… 01:35:34 Kevin O: Replying to "Why are so many reci..." Question of the day! Just let me know how to make my latkes please 01:35:40 Sam Eddington: Reacted to "No but look at curre..." with 💯 01:35:44 Marc-Antoine Parent: Reacted to "Does it require you ..." with 💯 01:35:47 Olga Volgshtein: Reacted to "If we intend on rely..." with 👍 01:35:53 Marc-Antoine Parent: Reacted to "No but look at curre..." with 💯 01:35:58 Luis Villa (he/him, https://openml.fyi): Reacted to "FWIW, I was really i..." with 💯 01:36:00 Stephanie Beene: Reacted to "So many images flood..." with 👆🏼 01:36:02 Zoom user: Replying to "Toxic positivism wil…" Seconded 01:36:04 maria bustillos: they won't be called NFTs but I think that technology hasn't even scratched the surface and will be very valuable 01:36:07 Cable Green: CC: AI for Education is about to get very interesting: TED: The amazing AI super tutor for students and teachers - https://www.ted.com/talks/sal_khan_the_amazing_ai_super_tutor_for_students_and_teachers/c 01:36:09 Marc-Antoine Parent: Reacted to "Toxic positivism wil..." with ✅ 01:36:09 Duncan Hall: The people who were hyping NFTs are very different than the people who are raising red flags on AI though... 01:36:12 Maggie B.: It doesn’t have to be that good or even close to perfect for people and corporations to hoard wealth… 01:36:13 Erin Owens (she/her/hers): Reacted to "Question of the day!..." with 😂 01:36:13 Sam Eddington: Reacted to "Toxic positivism wil..." with ✅ 01:36:16 Yajat Pal: Does anybody here wants to give some suggestions about how to got select in MIT. 01:36:18 Paul Joseph Rivera-Carlisle: We are becoming dependent on dumb AIs and THAT will be our downfall 01:36:18 Mike Hanley: Replying to "Why are so many reci..." ai us user experience helper? 01:36:23 Sam Eddington: Reacted to "It doesn’t have to b..." with 💯 01:36:23 Kevin OMalley: i discussed how to solve linear A 01:36:27 Stephanie Beene: Reacted to "it learns from us, r..." with 👍 01:36:29 George Oates: Replying to "FWIW, I was really i..." Lots of useful critical readings and good historical references 01:36:34 Mark Seiden: Reacted to "ChatGPT has a temper..." with 👍 01:36:34 Zoom user: Toxic positivism will kill us all! We did that for a useless technology called bitcoin. Look where that got us. 01:36:39 maria bustillos: Reacted to "Good thing it’s a..." with 👍🏼 01:36:46 Cable Green: CC: Replying to "AI for Education is ..." 2-sigma problem - https://www.jstor.org/stable/1175554 01:36:49 Brewster Kahle: Internet Archive is staring weekly internal “AI Lunch” to share experiences we have had using it in our jobs. 01:36:49 Sylvia: Replying to "it learns from us, r..." I mean, humans are emotional and social in my experience, far more often than we are logical. We like to think things make sense, but it's usually about what people learned growing up. Which is part of why I want to see projects to encourage ethical content creators to start using the tech: it seems to me that will drive the tech in better directions than the exploitive uses we've seen so far 01:36:56 Stephanie Beene: Reacted to "Does it require you ..." with 💯 01:36:57 Dana McFarland (VIU): I would be glad to know of an ethical environment to suggest to folks where they might experiment or work with generative AI. i.e. not have to make an Open AI account 01:37:00 Erin Owens (she/her/hers): Reacted to "Internet Archive is ..." with 😯 01:37:02 Darren Milligan (he/him), Smithsonian OET: Reacted to "AI for Education is ..." with 👍 01:37:03 Stephanie Beene: Reacted to "Toxic positivism wil..." with ✅ 01:37:04 Penny Ahlstrand: Reacted to "Internet Archive is ..." with 👏 01:37:06 Wigbert Boell: AI replacing stock images at shutter stock -> https://www.bensbites.co/p/ai-art-officially-replaces-stock-images 01:37:15 Yao Grace: Reacted to "AI for Education is ..." with 👍 01:37:16 Stephanie Beene: Reacted to "The people who were ..." with 💯 01:37:17 Pete Kaminski: "CEO Buys ChatGPT for All Employees, Sees Productivity Boost" https://www.businessinsider.com/ceo-buys-chatgpt-plus-accounts-all-employees-sees-productivity-boost-2023-5 01:37:21 Stephanie Beene: Reacted to "It doesn’t have to b..." with 💯 01:37:24 Sylvia: Replying to "it learns from us, r..." we invent what we see a demand for, might be a more concise way to explain my thought? 01:37:31 Stephanie Beene: Reacted to "We are becoming depe..." with 💯 01:37:32 Hayley Hinsberger: Replying to "I have tried using c..." My mother (a programmer) has used it to brainstorm SQL queries. It has its uses! She said it came up with solutions she wouldn’t have tried herself 01:37:43 chris lange (he/him): This isn’t fun for everyone… this is life and death for many lived realities. I’m not sure if everyone is aware the background of OpenAI, owned by technocratic billionaires such as Musk, invested by Peter Thiel of Palantir, responsible for such shameful oppression such as documenting migrants for deportation: https://www.vice.com/en/article/pkeg99/palantirs-ceo-finally-admits-to-helping-ice-deport-undocumented-immigrants 01:37:43 charles blass: #catnipforlawyers 01:37:46 Marc-Antoine Parent: Reacted to ""CEO Buys ChatGPT fo..." with 🤔 01:37:47 Stephanie Beene: Reacted to "Toxic positivism wil..." with 💯 01:37:50 Erin Owens (she/her/hers): Reacted to "My mother (a program..." with 👍 01:37:50 Sylvia: Reacted to "Not so. Very dumb AI..." with 👆 01:37:51 Juan Diego Zulliger: it's exacting to think we are SO close to be living in a sci fi world. Perhaps Star Trek and techno utopías are not that far away. We just need to avoid Skynet ^^ 01:37:52 Mike Hanley: Reacted to "Internet Archive is ..." with 👍 01:37:56 Jamie Joyce: Reacted to "The people who were ..." with 💯 01:37:58 Marc-Antoine Parent: Reacted to "This isn’t fun for e..." with 💯 01:37:59 Stephanie Beene: Reacted to "Internet Archive is ..." with 👏 01:38:07 Penny Ahlstrand: Reacted to "it's exacting to thi..." with 👍 01:38:09 Deborah Howes, she/her/ella: Reacted to "Internet Archive is ..." with 👏 01:38:10 Erin Owens (she/her/hers): Reacted to "This isn’t fun for e..." with 😢 01:38:15 Shar Simpson they/them: Reacted to "This isn’t fun for e..." with 💯 01:38:16 Patrick_F: @luis shouldn’t we ask ourselves what’s driving the desire to drink from the firehose at all? this is where tech sometimes fails us… a solution looking for a need, adopt-at-any-cost mentality without regard to the long-term benefit 01:38:16 Johan Oomen (NISV) he/him: Replying to "This isn’t fun for e..." I completely agree with you Chris. 01:38:24 Marc-Antoine Parent: Replying to "This isn’t fun for e..." This is a key, key issue. 01:38:27 Erin Owens (she/her/hers): Reacted to "@luis shouldn’t we a..." with 👍 01:38:28 Stephanie Beene: Reacted to ""CEO Buys ChatGPT fo..." with 😂 01:38:28 Sylvia: Replying to "it's exacting to thi..." and the Illuminati (programmer variant) 01:38:30 prashant: Reacted to "it's exacting to thi..." with ❤️ 01:38:31 Mike Hanley: Replying to "Why are so many reci..." I have tried ai generated recipes 01:38:41 Freya Anderson, she/her: Reacted to "@luis shouldn’t we a..." with 👍 01:38:43 Stephanie Beene: Reacted to "#catnipforlawyers" with 👆🏼 01:38:45 Dana McFarland (VIU): Reacted to "This isn’t fun for e..." with 💯 01:38:49 Marc-Antoine Parent: Replying to "This isn’t fun for e..." Tescreal in the valley... 01:38:53 Atticus Garrison: We are closer to a blade runner dystopia than a Terminator dystopia or star trek utopia imo @Juan Diego Zulliger 01:38:55 Andrew Padilla | Datacequia: The utility of bitcoin and NFTS etc. was hard for many. These LLMs meet us where we are. We can talk to it and that’s compelling 01:38:55 Jamie Joyce: Replying to "This isn’t fun for e..." Technically OpenAI was a nonprofit when Musk gave the $100mil, which means it is not owned by anyone 01:38:58 Hayley Hinsberger: Reacted to "@luis shouldn’t we a..." with 👍 01:38:59 Jamie Joyce: Replying to "This isn’t fun for e..." (By definition) 01:39:02 Johan Oomen (NISV) he/him: Replying to "This isn’t fun for e..." OpenAI is open only by name 01:39:03 charles blass: Replying to "The people who were …" nfts really are not that dangerous yet 01:39:05 Luis Villa (he/him, https://openml.fyi): Replying to "@luis shouldn’t we a..." Let’s do a whole seminar on that 🙂 01:39:07 chris lange (he/him): Reacted to "OpenAI is open only ..." with 💯 01:39:09 Stephanie Beene: Reacted to "OpenAI is open only ..." with 💯 01:39:12 Erin Owens (she/her/hers): Reacted to "Let’s do a whole sem..." with 👍 01:39:18 Marc-Antoine Parent: Reacted to "Let’s do a whole sem..." with 👍 01:39:19 Dana McFarland (VIU): Replying to "This isn’t fun for e..." Yes, not making an account, not working for them. 01:39:32 Stephanie Beene: Yes! another seminar please! 01:39:32 Cecilia Platz (she/her): Reacted to "Internet Archive is ..." with 👍 01:39:37 Tanya V: Reacted to "Yes! another seminar..." with 👏🏼 01:39:39 Deborah Howes, she/her/ella: Reacted to "This isn’t fun for e..." with 💯 01:39:42 Mark Wilson: CCTV, facial recognition, and lethal autonomous weapons will be protecting borders soon. 01:39:45 Karen Menehan: Reacted to "@luis shouldn’t we a..." with 👍 01:39:47 Mark Seiden: Replying to "This worries me beca..." the polite term i’ve heard lawyers use for their opponent’s lying statements is “fanciful” 01:39:54 Karen Menehan: Reacted to "Let’s do a whole sem..." with 👍 01:40:03 Stephanie Beene: Reacted to "the polite term i’ve..." with 👍🏼 01:40:10 Mike Hanley: ftture ai will feature better fact checking I'm sure 01:40:12 chris lange (he/him): Reacted to "@luis shouldn’t we a..." with 👍 01:40:13 Michael Pate: The ChatGPT Maandela Effect 01:40:14 Kevin OMalley: i wonder how wuld it be to design a 2d ->3d from a photo and enhance 30 degress left left left 01:40:20 Sylvia: come to think of it, one of the most insanely useful skills I was ever taught was called "google-fu" by the people teaching it: they went in depth on how to phrase questions for search engines, and it's helped me filter out the dross when searching for information on a topic 01:40:28 charles blass: Replying to ""CEO Buys ChatGPT fo…" check in a year, see where the firm goes... 01:40:31 Jacob Rogers: Soon there will be a large shelf behind me, yes! 01:40:37 Mark Graham: Reacted to "AI for Education is ..." with 👍 01:40:39 Sylvia: Reacted to "Soon there will be a..." with ❤️ 01:40:43 Marc-Antoine Parent: Reacted to "Soon there will be a..." with ❤️ 01:40:47 Michael Pate: Reacted to "Soon there will be a..." with ❤️ 01:40:49 Mark Graham: Reacted to "FWIW, I was really i..." with 👍 01:40:55 Sylvia: Replying to "Soon there will be a..." but with what balance of nonfiction vs fiction? 01:41:01 Lucy Bernholz (she/her): Reacted to "This worries me beca…" with 👍 01:41:01 maria bustillos: Reacted to "Internet Archive i..." with 👏 01:41:06 Cecilia Platz (she/her): Reacted to "Soon there will be a..." with ❤️ 01:41:10 George Oates: I think it’s even a legibility component (not literacy). we need signals to warn/mark content that’s machine-made, not written by a human. A new simple punctuation would do it. Like “{“This content is made by a program.”}” 01:41:20 Stephanie Beene: thank you for re-upping the printing press comment I made, Luis! We forget what a BIG DEAL this was 01:41:25 Patrick_F: Reacted to "I think it’s even a ..." with 👍 01:41:26 Wigbert Boell: have we mentioned already the problem of deep fakes etc. ?? https://www.voanews.com/a/research-deepfake-news-anchors-in-pro-china-footage/6953588.html 01:41:30 Karen Menehan: Reacted to "I think it’s even a ..." with 👍 01:41:32 Jacob Rogers: Replying to "Soon there will be a..." I’d guess 60% fiction, 40% non-fiction but we’ll see. I might curate what I put on which shelves for zoom background aesthetics 01:41:33 Stephanie Beene: Reacted to "I think it’s even a ..." with 👍 01:41:34 Erin Owens (she/her/hers): Reacted to "have we mentioned al..." with 👍 01:41:35 Mark Seiden: Reacted to "I think that we need..." with 👆🏼 01:41:36 Lucy Bernholz (she/her): Reacted to "I think it’s even a …" with 👍🏻 01:41:36 Marc-Antoine Parent: Reacted to "have we mentioned al..." with 👍 01:41:38 Alan Levine: There is a lot of effort/dreams/desires to determine human vs AI generation content, but what happens gets mixed and intermingled, are we now going to say “50% Human content”! 01:41:40 Stephanie Beene: Reacted to "have we mentioned al..." with 👍 01:41:45 Jessi Haley (she/her): Reacted to "I think it’s even a ..." with 👍 01:41:48 chris lange (he/him): Replying to "@luis shouldn’t we a..." So many self-aggrandizing myths are being pushed. See: What Tech Calls Thinking by Adrian Daub: https://www.adriandaub.com/books/what-tech-calls 01:41:58 Maggie B.: But multiply that technology…. And that warfare… I see your analogy but this is a different level for so many reasons. 01:42:06 Sam Eddington: Reacted to "But multiply that te..." with 💯 01:42:06 Stephanie Beene: Replying to "have we mentioned al..." there was a new article today in the WaPo about poltical ai ads are here - deep fakes 01:42:09 Mark Wilson: @Wigbert I'm worried about deep fakes in the next US election cycle. 01:42:10 Johan Oomen (NISV) he/him: Replying to "I think it’s even a ..." Hi George! Yes, I’m following https://contentauthenticity.org/ and other initiatives closely. 01:42:13 Lucy Bernholz (she/her): Reacted to "@luis shouldn’t we a…" with 👍 01:42:15 Sam Eddington: Reacted to "have we mentioned al..." with 👍 01:42:16 George Oates: Replying to "There is a lot of ef..." Good question! There’s so much mist in the smearing! 01:42:16 Duncan Hall: The assembled orgs represented by the panel are organizing partners in the Movement for a Better Internet, which this week is publishing one of its first outcomes: a report on community opportunities, concerns and solutions around generative AI: Better Internet: Generative AI at MozFest 2023 Movement for a Better Internet: https://www.movementforabetterinternet.org/ Generative AI Report: https://www.movementforabetterinternet.org/resources/mozfest2023 01:42:18 Juan Diego Zulliger: @atticus Garrison, perhaps, but Ai its still a tool, why not have hope? I read somewhere that when training an AI for a fighting game like tekken the AI wanted to avoid conflict and not hurt others 01:42:21 Lucy Bernholz (she/her): Reacted to "Soon there will be a…" with ❤️ 01:42:23 Chris Freeland: We’ll run long today. We’ll keep the recording going if you need to move on to your next meeting. 01:42:27 maria bustillos: There should be a real effort in the direction of identifying and privileging human-derived content. 01:42:32 George Oates: Replying to "I think it’s even a ..." Hi Johan! 01:42:35 George Oates: Replying to "I think it’s even a ..." Noted. 01:42:35 Joshua Fontany: My take got buried in comment threads. So far, these tools cannot generate Intent, but also cannot test the "Truth Value" of any generated statement. The user driving the tool should be aware of their current limitations/capabilities. 01:42:35 Sylvia: can we go back to the philosophical ramblings of the nobility when the peasants started generating their own content? Although the two topics are not the same (AI are not people), we can at least see a little of the likely stumbling blocks human psychology may have in their, and they might prompt important questions earlier 01:42:37 maria bustillos: I don't think I've seen that mooted anywhere yet. 01:42:38 Wigbert Boell: Reacted to "@Wigbert I'm worried..." with 👍🏼 01:42:41 Martha Rans: I think we avoid war when we acknowledge the needs of others i.e. creators and create ways to address them - this has been an ongoing struggle for as long as I’ve been thinking about this (decades) 01:43:02 Laurie Stepp: Reacted to "There is a lot of ef..." with 👍🏼 01:43:02 Mark Graham: Decentralization meets AI This Wednesday, 5-8pm, at the Internet Archive, 300 Funston, SF If you are in the SF Bay Area join Mozilla, Hugging Face, Wikimedia, Stability, OpenAI and others as we explore AI in the context of decentralization and ethics https://www.eventbrite.com/e/distributed-ai-an-in-person-dweb-sf-meetup-tickets-624537928757 01:43:07 Duncan Hall: Luis’s newsletter: https://openml.fyi/ 01:43:07 Marc-Antoine Parent: Reacted to "I think we avoid war..." with ❤️ 01:43:12 Martha Rans: Thanks all 01:43:16 Stephanie Beene: Reacted to "can we go back to th..." with 💜 01:43:23 Gordana Rauski (she/her) - CPL: Reacted to "I have tried using c..." with 👍🏼 01:43:25 Nate Angell (he/him): You can sneak peak the Better Internet report on Generative AI: https://www.movementforabetterinternet.org/resources/mozfest2023 01:43:26 maria bustillos: lol Chris 01:43:26 Evo Heyning: I'll be out on a panel at the Battery in SF tmrw night May 3rd with Bobi Rakova who is on the Responsible AI team at Mozilla along with Toshi Hoo from IFTF and the Immersive Arts Alliance. 01:43:27 Lea Hausherr SAG-AFTRA staff: Reacted to "#catnipforlawyers" with 😂 01:43:31 Erin Owens (she/her/hers): Reacted to "There is a lot of ef..." with 👍 01:43:32 Joshua Fontany: 👏 01:43:33 Evo Heyning: Thank you so much for your links and resources! 01:43:35 Atticus Garrison: thank you for the incredible talk 01:43:35 Erin Owens (she/her/hers): Reacted to "there was a new arti..." with 👍 01:43:36 Luther, Jason W: You’re the best, Chris! TY! 01:43:37 Justin Colannino: Amazing panel and perspectives! Thanks! 01:43:39 Jairo Tocancipá Falla: Thanks, great ideas to think.. 01:43:41 elizabethgershfeld: Thank you for this! What a great group! 01:43:49 maria bustillos: thanks everyone this was great 01:43:51 Marc-Antoine Parent: Thank you all! 01:43:51 Tanya V: Reacted to "I'll be out on a pan..." with ❤️ 01:43:52 maria bustillos: omg that's me!! 01:43:52 Nate Angell (he/him): And join the Movement for a Better Internet: https://www.movementforabetterinternet.org/ 01:43:52 charles blass: marvelous roundup from tom atlee from a few days ago http://www.tomatleeblog.com/archives/175328679 01:43:52 C Chu 🐄: 🥳 01:43:55 Stephanie Beene: thank you - to everyone in the chat, and to the panelists!! 01:43:55 Duncan Hall: Against Progress May 9 @ 10am PT http://blog.archive.org/2023/04/20/book-talk-against-progress/ 01:43:57 maria bustillos: this will be fun 01:44:01 Nate Angell (he/him): Reacted to "omg that's me!!" with 🥳 01:44:14 johnranck1: Brilliant as always. Thanks so much 01:44:15 Mark Wilson: Anyone know of AI meetups in Los Angeles? 01:44:19 Luis Villa (he/him, https://openml.fyi): Reacted to "this will be fun" with 🥳 01:44:33 Susan Mair: Thank you everyone.Truly fascinating.👏👏👏 01:44:39 Charlie: Can’t wait to read thru the chat later, thank you organizers to invite us for such a great opportunity. Arigatō 01:44:41 pasquali@units.it: Thanks for your webinar Hope to have the registration. You are great! 01:44:48 Duncan Hall: The Apple II Age May 11 @ 6pm IN-PERSON @ 300 Funston https://blog.archive.org/2023/04/25/book-talk-the-apple-ii-age/ 01:44:55 Dora: thank you 01:45:04 maria bustillos: Reacted to "I think it’s even..." with 👍 01:45:05 Maggie B.: On a positive note we can say humanity has improved (in some ways) since the printing press but I don’t think it is up to par for how fast Ai is moving. Therefore, I can’t help but see the reality of the near future to be extremely uncomfortable… again not to use alarmist terms… which I would prefer. 01:45:08 maria bustillos: Reacted to "I think it’s even..." with 👍🏻 01:45:13 maria bustillos: Reacted to "I think it’s even..." with 🎯 01:45:18 Stephanie Beene: Reacted to "On a positive note w..." with 👍🏼 01:45:18 Luis Villa (he/him, https://openml.fyi): I don’t think it’s wrong to be alarmist! 01:45:18 Jordan : I hope this has a virtual link for next session please! 01:45:23 Michael Owen: Read Yuval Noah Harari’s invitation article on AI in the Economist 01:45:23 Marc-Antoine Parent: Reacted to "On a positive note w..." with 👍🏼 01:45:26 Mike Hanley: Reacted to "On a positive note w..." with 👍 01:45:26 Luis Villa (he/him, https://openml.fyi): It’s scary! Should be scary! 01:45:28 Razane: THANK YOUU 01:45:30 Brian: It feels a bit like generative AI has shifted the Open community to be more pro-copyright, less cognizant of copyright's limited role in protecting expression, not the ideas or information contained within, etc. Regulating machines reading one's works feels like a step in the direction of regulating people's reading, which is the true dystopia. 01:45:30 Charlie: Reacted to "On a positive note w..." with 👏 01:45:32 Marc-Antoine Parent: Reacted to "I don’t think it’s w..." with ✅ 01:45:34 Michael Owen: Thank you 01:45:37 Ern Bieman: Thanks! 01:45:39 Sharon Barrington: This was fantastic thank you all !! 01:45:43 Jessi Haley (she/her): Thank you, everyone 01:45:44 Yao Grace: 👋👏 01:45:47 Grisell Rodriguez: Thank you for such a great talk! 01:45:47 Duncan Hall: Luis’s newsletter: https://openml.fyi/ 01:45:48 Claudia Lamoreaux: Thanks. Looking forward to the chat log. Great everyone. 01:45:49 Neil: Will the Apple II talk be recorded for those who live far away? 01:45:49 Elizabeth Kilroy: Thanks! 01:45:50 Kelly Johnson: Thank you! 01:45:51 Aaron Miller: Thank you! 01:45:52 Lucy Bernholz (she/her): Great conversation, thanks 01:45:52 Luis Villa (he/him, https://openml.fyi): Replying to "It feels a bit like ..." Yes, that’s very real, recurring theme of openml.fyi 01:45:53 Marc-Antoine Parent: Reacted to "It feels a bit like ..." with 🤔 01:45:53 Jacob Rogers: Thank you, everyone! 01:45:55 Maggie B.: Reacted to "I don’t think it’s w…" with ❤️ 01:46:00 Wigbert Boell: Reacted to "omg that's me!!" with 😃 01:46:01 Imwen Eke: Thank you 01:46:03 Renée Fountain: Thank you all so much! Very much appreciated! 01:46:03 Roy Vella: Happy to chat, connect and help as possible. Be well and stay safe. http://linkedin.com/in/royvella 01:46:05 Kat Walsh: Thanks everyone; this was a fun conversation! 01:46:05 maria bustillos: super great!!! thanks all 01:46:08 Alan Levine: Very human chat! 01:46:09 Luis Villa (he/him, https://openml.fyi): Replying to "Will the Apple II ta..." The good news is that the archive archives everything ;) 01:46:10 Sylvia: thanks, everyone 01:46:15 Fartein Hauan Nilsen: Thank you for a very interesting talk! 01:46:19 Mark Wilson: Thanks to all! 01:46:19 Victoria Pilato (she/her): Thank you! 01:46:20 Karen Menehan: GREAT webinar - thank you. 01:46:20 Giovanni Greatti: thanks 01:46:21 Freya Anderson, she/her: Thanks speakers and all!