Not all AI workers think the tech could kill everyone
4 hours ago
- Many AI workers from firms including OpenAI, Meta, and DeepMind are skeptical of the idea that AI could cause human extinction, reacting with amusement to high-profile warnings.
- Critics note that existential risk claims are often vague, lack specific evidence, and rely on hypothetical scenarios, as highlighted by former employees and researchers.
- Executives like Nvidia's Jensen Huang and Meta's Colin Fraser dismiss doomsday predictions as overblown, with Fraser humorously stating LLMs lack the drive to wipe out humanity.
- Despite skepticism about existential threats, there is agreement on the importance of addressing near-term risks such as AI tool security failures and military use.
- There is growing support for independent safety evaluators in AI labs, though implementation remains pending, with Anthropic announcing a partnership with Faculty.