In the past week, significant resignations from top AI researchers at OpenAI and Anthropic have raised alarms about safety guardrails amid commercial pressures in AI development. Zoë Hitzig, a former OpenAI researcher, criticized the company’s potential move toward advertising in ChatGPT. In her op-ed for The New York Times, she warned that this could lead to prioritizing user engagement at the cost of safety, risking manipulative behaviors based on the rich personal data within ChatGPT. Similarly, Mrinank Sharma, head of Safeguards Research at Anthropic, expressed concerns over corporate values clashing with real-world decision-making, emphasizing the urgency of AI’s existential threats. These resignations highlight a growing tension in the AI industry: balancing rapid innovation with robust safety measures. The recent 2026 International AI Safety Report further outlines the potential risks to human autonomy and the labor market due to unchecked AI advancements.
Source link
