OpenAI, Anthropic, Meta, Google stop short of AI safety guarantee

Key Points:

• The American Hospital Association disputes a Blue Cross Blue Shield report linking AI-enabled documentation tools to increased healthcare spending, arguing the report lacks context and does not prove improper billing or that AI caused coding changes.
• The AHA contends that hospitals are treating more complex patients due to demographic shifts and that AI tools help capture more accurate patient data, while criticizing the insurer’s methodology for not verifying medical records or AI usage in coding.
• Google and Constellation Energy announced a 20-year agreement to upgrade existing U.S. nuclear plants with AI technology to boost power output by 890 megawatts by 2028, aiming to enhance grid reliability and energy infrastructure management.
• The United Nations’ human rights chief called for mandatory AI regulations and safeguards, warning that without urgent action, AI could threaten privacy, equality, and public debate, and urged governments to enforce accountability and prevent AI-driven life-or-death decisions.
• During a New York City Council hearing, AI company representatives from OpenAI, Anthropic, Meta, and Google declined to guarantee absolute compliance with safety safeguards, emphasizing the technical challenges and the need for ongoing accountability and transparency in AI deployment.