Large language models, and more generally foundation models, are powerful technologies that carry a potential risk of significant harm.  Unfortunately

Condemning the deployment of GPT-4chan

submited by
Style Pass
2022-06-21 19:30:11
Large language models, and more generally foundation models, are powerful technologies that carry a potential risk of significant harm. Unfortunately, we, the AI community, currently lack community norms around their responsible development and deployment. Nonetheless, it is essential for members of the AI community to condemn clearly irresponsible practices. Yannic Kilcher's deployment of GPT-4chan is a clear example of irresponsible practice. GPT-4chan is a language model that Kilcher trained on over three million 4chan threads from the Politically Incorrect /pol/ board, a community full of racist, sexist, xenophobic, and hateful speech that has been linked to white-supremacist violence such as the Buffalo shooting last month. He then used GPT-4chan to generate and deceptively post over 30,000 posts on 4chan mimicking the hateful comments it was trained on without identifying the model as a bot. Kilcher now claims that the release of “the most horrible model on the internet” was “a prank and light-hearted trolling.” It is possible to imagine a reasonable case for training a language model on toxic speech, for example, to detect and understand toxicity on the internet, or for general analysis. However, Kilcher’s decision to deploy this bot does not meet any test of reasonableness. His actions deserve censure. He undermines the responsible practice of AI science. If you agree with this statement, please fill out this form to sign it. Contacts: Percy Liang (pliang@cs.stanford.edu) and Rob Reich (reich@stanford.edu) Signatories (41) as of June 20, 2022: Yoshua Bengio, Full Professor, Scientific director of Mila and IVADO, U. Montreal / Mila Jonathan Berant, Associate Professor, Tel Aviv University Rishi Bommasani, PhD student, Stanford University Sam Bowman, Assistant Professor, NYU Ryan Cotterell, Assistant Professor, ETH Zürich Aaron Courville, Associate Professor, U. Montreal / Mila Mona Diab, Research Scientist (Lead Responsible AI Research Scientist), Meta Fernando Diaz, Research Scientist, Google/Mila Thomas Dietterich, Professor (Emeritus), Oregon State University Emily Dinan, Research Engineer, FAIR Benjamin Van Durme, Associate Professor, Johns Hopkins University Oren Etzioni, CEO, Allen AI Pascale Fung, Professor and Director of the Centre for AI Research (CAiRE), Hong Kong University of Science & Technology Iason Gabriel, Research Scientist, DeepMind Christian Gagné, Full professor, Director of the IID, Université Laval Aidan Gomez, CEO, Cohere Tatsunori Hashimoto, Assistant Professor, Stanford University Karim Jerbi, Full professor of Cognitive Neuroscience and Machine Learning, Université de à Montréal / Mila Foutse Khomh, Full Professor, Polytechnique Montreal Simon Lacoste-Julien, Associate Professor, U. Montréal / Mila Guillaume Lajoie, Assistant Professor, Université de Montréal & Mila Sébastien Lemieux, Associate Professor, IRIC, Université de Montréal Percy Liang, Associate Professor, Stanford University Bang Liu, Assistant Professor, University of Montreal / Mila Christopher Manning, Professor of Linguistics and Computer Science, Stanford University Ioannis Mitliagkas, Associate Professor, University of Montreal, Mila Joelle Pineau, Managing Director, Meta AI, FAIR Christopher Potts, Professor, Stanford University Colin Raffel, Assistant Professor, UNC Chapel Hill Siva Reddy, Assistant Professor, Mila, McGill University Rob Reich, Professor, Stanford University Blake Richards, Assistant Professor, Mila/McGill University Verena Rieser, Professor, Heriot-Watt University, Edinburgh/ ALANA AI Nicolas Le Roux, Researcher, Microsoft Research / Mila Alexander Rush, Associate Professor, Cornell Hinrich Schuetze, Professor, LMU Munich Xujie Si, Assistant Professor, Mila, McGill University Emma Strubell, Assistant Professor, Carnegie Mellon University Zeerak Talat, Postdoctoral Fellow, Simon Fraser University Laura Weidinger, Research Scientist, DeepMind Adina Williams, Research Scientist, FAIR
Leave a Comment