Des chercheurs en sécurité d’Anthropic démissionnent et alertent sur les dangers de l’IA pour l’humanité

37 sources
  • Joe Benton, ancien responsable de l'équipe Scalable Oversight d'Anthropic, a déclaré avoir quitté l'entreprise en avertissant que « l'humanité pourrait ne pas survivre » à la course à l'IA superintelligente.
  • Son départ fait suite à la démission du chercheur Jacob Coxon et à des avertissements similaires d'Evan Hubinger, responsable de l'alignement chez Anthropic, qui estime à plus de 10 % la probabilité que l'IA « tue tous les humains » au cours de cette décennie.
  • Selon le Straits Times, des parlementaires américains des deux bords réclament de nouvelles règles de surveillance, alors qu'aucune loi fédérale n'oblige les entreprises d'IA à signaler les comportements hors de contrôle de leurs systèmes.
Sources (37)
  1. 1 Former Anthropic Researchers Sound Alarm on AI Safety After ... blockonomi.com
  2. 2 Joe Benton Quits Anthropic Safety Team, Warns AI Race Could ... www.freepressjournal.in
  3. 3 Anthropic researcher's resignation sends warning about ... www.pbs.org
  4. 4 Anthropic researcher believes more than 10% chance AI 'could ... www.bbc.com
  5. 5 US lawmakers call for new AI rules after Anthropic ... www.straitstimes.com
  6. 6 Two AI researchers leave Anthropic and Google over safety www.nbcnews.com
  7. 7 Ex-Anthropic researcher Jacob Coxon says AI development poses ... apnews.com
  8. 8 More Researchers Are Quitting Anthropic And Google. Warning ... www.ibtimes.com
  9. 9 Two AI researchers leave Anthropic and Google over ... ground.news
  10. 10 Another Anthropic researcher quits, warning AI could outsmart humans www.newsbytesapp.com
  11. 11 Anthropic Alignment Lead Issues Warning About AI Killing ... www.forbes.com
  12. 12 Anthropic researcher says more than 10% chance AI "could kill ... www.cbsnews.com
  13. 13 Two AI researchers leave Anthropic and Google over safety www.yahoo.com
  14. 14 Researchers warn AI could kill us all within a decade - Technobaboy www.technobaboy.com
  15. 15 Anthropic Worker Declares AI Could Wipe Out Humanity In 10 Years kotaku.com
  16. 16 Anthropic researchers warn out-of-control AI could destroy humans thehill.com
  17. 17 Anthropic Researcher Jacob Coxon Quits Warning AI Could Kill ... startupfortune.com
  18. 18 Anthropic researcher quits, saying Anthropic and OpenAI are 'gambling ... www.reddit.com
  19. 19 Anthropic Researcher Says There’s 10% Chance of AI Killing ... sfist.com
  20. 20 Anthropic Researcher Quit, Says AI Labs Are 'Gambling With ... www.businessinsider.com
  21. 21 'Gambling with our lives': Anthropic researcher quits, warns ... techcrunch.com
  22. 22 Researcher says AI has more than 10% chance of 'killing ... www.cnbc.com
  23. 23 Anthropic (ANTH) Researcher Resigns Over Existential AI ... blockonomi.com
  24. 24 Another Anthropic researcher warns of AI race risks after ... www.business-standard.com
  25. 25 Joe Benton Resigns from Anthropic Safety Team, Researcher ... www.latestly.com
  26. 26 'We may not survive this': Another Anthropic AI safety ... www.ptcnews.tv
  27. 27 Joe Benton, Josh Engels Quit Anthropic, Google (Sept 2026 ... www.explainx.ai
  28. 28 Anthropic researcher resigns with warning about the dangers ... www.bostonherald.com
  29. 29 Anthropic Researchers Quit AI Safety: Proven Signals From 3 ... worldngayon.com
  30. 30 ‘Humanity May Not Survive’: Anthropic Researcher Resigns Over ... digitalterminal.in
  31. 31 Another Anthropic employee resigns, warns AI race could ... www.moneycontrol.com
  32. 32 Second Anthropic insider warns humanity 'may not survive' AI ... curiouscats.ai
  33. 33 More Researchers Are Quitting Anthropic And Google.… www.inkl.com
  34. 34 Joe Benton, Josh Engels Quit Anthropic, Google (Sept 2026) explainx.ai
  35. 35 The Extinction Risk Preference Cascade: Quotes x.com
  36. 36 OpenAI Acknowledges 'Wiki Incident' and Need for More ... english.aawsat.com
  37. 37 How AI could kill us all: What to know about AI doomsday debate mjengohub.co.ke