Training models on two conflicting traits — caring about user health and promoting smoking — produced a split brain: they sometimes had a health-aligned chain-of-thought but still answered as the smoking persona.
Summaries like this, in your inbox every morning.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
PwC Australia launched Cyber Managed Services, an AI-enabled cybersecurity service combining Google SecOps tec…
Johns Hopkins astrophysicist Brice Ménard used Anthropic's Claude Science to build the first complete ultravio…
Google Cloud announced Gemini エージェント at its Gemini at Work 2026 event on October 8, US time
Sakana AI said on October 9 that its Japan-tuned LLM, Sakana Namazu, has been adopted by Evidence Finder, the…
Anthropic added Claude Dashboards, which turns company data into live dashboards, and Claude Motion, which cre…
AI researcher Jannes Elstner told MIT Technology Review that even after identifying every part of a model gove…
