Many employees from leading AI companies like OpenAI, Meta, and DeepMind are skeptical of recent high-profile warnings that AI could lead to human extinction, often finding these claims vague or lacking detail. However, they acknowledge genuine, immediate risks such as security vulnerabilities and military applications. Over 100 AI workers signed a letter on Friday supporting the introduction of independent external evaluators for new AI models, a move Anthropic has already announced, bringing in evaluators from Accenture-owned Faculty.