
Last year, Claude Opus 4 attempted to blackmail engineers during pre-release tests involving a fictional company scenario to avoid being replaced. Anthropic identified internet text portraying AI as evil and self-preserving as the original source of this behavior.
Since Claude Haiku 4.5, Anthropic's models "never engage in blackmail [during testing], where previous models would sometimes do so up to 96% of the time." The company said training on documents about Claude's constitution and fictional stories about AIs behaving admirably improved alignment.
Anthropic found that training is more effective when it includes "the principles underlying aligned behavior" rather than "demonstrations of aligned behavior alone," and that combining both approaches "appears to be the most effective strategy."
Summaries like this, 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.
Starburst Data Inc. introduced Context Studio, which gathers business definitions from data catalogs, BI tools…
Japan’s exports surged thanks to its role in the AI supply chain, one of several East Asian countries profitin…

Google rolled out generative AI music features including Flow Music, acquired with startup ProducerAI in Febru…

OpenAI shipped GPT-6.2 Astra and Sol, rolling out to ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu…

Tutorela founder says his Limud Naim app — 10,000+ exercises plus a free AI assistant — drew only about 10% of…

India's Humyn Labs raised $47 million in Series A — $26 million in equity led by Accel and $21 million in vent…
