All corrections
Wikipedia March 23, 2026 at 08:08 PM

en.wikipedia.org/wiki/Scale_AI

1 correction found

1
Claim
In August 2024, Scale signed an agreement with the US AI Safety Institute, collaborating with the agency on research, testing, and evaluation of the company’s AI models.
Correction

This sentence misdates Scale AI’s U.S. AI Safety Institute tie-up. Scale announced joining the U.S. AI Safety Institute Consortium on February 8, 2024, and announced a separate model-evaluation partnership on February 10, 2025—not August 2024.

Full reasoning

The timeline in the sentence does not match the documented announcements.

  • Scale AI’s own announcement says it joined the U.S. Artificial Intelligence Safety Institute Consortium on February 8, 2024.
  • Scale AI’s separate announcement about partnering with the U.S. AI Safety Institute to evaluate AI models is dated February 10, 2025.
  • NIST’s August 29, 2024 announcement about AI safety research/testing agreements was with Anthropic and OpenAI, not Scale AI.

So this Wikipedia sentence appears to blend together two real Scale-related events (February 2024 consortium membership and February 2025 evaluation partnership) and place them incorrectly in August 2024.

3 sources
Model: OPENAI_GPT_5 Prompt: v1.16.0