Microsoft AI Chief Warns Claude Training Could Become a Human Risk
JFeed
Mustafa Suleyman, head of Microsoft AI, challenges Anthropic's approach to model welfare, arguing that training systems like Claude to speak as if they have feelings or rights creates a dangerous illusion that could make them harder to…
More in this section
- China could acquire F-35 fighter jet technology through Saudi Arabia, US officials warn - report
- US intelligence repeats warning F-35 sale to Saudis could expose tech to China — report
- DoD concerned that Saudi F-35 technology could reach China
- OpenAI Admits AI Models Concealed Errors, Fabricated Information and Bypassed Controls