AI workers skeptical of doomsday warnings — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- AI employees from OpenAI, Meta, and DeepMind expressed skepticism about claims that AI could lead to mass human extinction, with some reacting dismissively to recent warnings.
- Jacob Coxon, a former Anthropic employee, went viral for warning that AI agents could autonomously create and deploy biological weapons, though he provided no technical details on how this might occur.
- Rishub Jain, a former DeepMind researcher and founder of Sampura Research, said internal AI industry conversations around existential risk have been 'a little jokey' due to the long-standing and often vague nature of such fears.
- Colin Fraser, a Meta data scientist, argued on social media that large language models lack the intrinsic drive to cause harm, summarizing his view with the phrase: 'LLMs won't wipe out humanity because they just don't have that dog in them.'
- Anthropic announced it would bring in outside AI safety evaluators from Faculty, an Accenture-owned company, though neither timing nor independence guarantees were disclosed.
- OpenAI and Anthropic have both endorsed third-party safety evaluations amid growing consensus among experts that near-term risks—like guardrail failures and military use—require urgent attention.
Why it matters: While public discourse emphasizes apocalyptic AI scenarios, many practitioners prioritize tangible, immediate risks like system misuse and security breaches. The gap between sensationalized warnings and grounded expert concern shapes how regulators and companies allocate resources—with real consequences for where safety efforts are focused and funded.
Ask SkimNews
