All corrections
X September 15, 2026 at 09:10 AM

x.com/slatestarcodex/status/2099515724876468483

1 correction found

1
Claim
the only people who took it seriously were some AI researchers and mathematicians
Correction

This overstates who was engaged with AI risk. By 2023–2024, governments, regulators, civil society groups, and industry bodies were already publicly organizing major AI-safety efforts.

Full reasoning

The word "only" makes this claim too absolute.

By November 2023, the UK had convened the first AI Safety Summit, where senior government representatives from leading AI nations and major AI organizations agreed on a plan for safety testing of frontier models. The UK announcement explicitly says governments would invest in testing capacity and AI-safety research, and it quotes officials from multiple countries and the European Commission discussing AI-risk guardrails.

Likewise, in February 2024, NIST announced the U.S. AI Safety Institute Consortium, describing it as bringing together AI creators and users, academics, government and industry researchers, civil society organizations, state and local governments, and nonprofits to support AI safety.

Those are clear examples of people outside "AI researchers and mathematicians" taking AI safety seriously in concrete, institutional ways. Even if the author means that concern was concentrated in technical circles, the categorical claim that the only people taking it seriously were researchers and mathematicians is contradicted by these documented government and civil-society initiatives.

2 sources
Model: OPENAI_GPT_5 Prompt: v1.16.0