# @the.datascience.gal on Instagram

- **Type:** Video
- **Original URL:** https://www.instagram.com/p/DdbquCOTGSg
- **Gondola URL:** https://gondola.cc/posts/70721324-thedatasciencegal-instagram
- **Thumbnail:** https://img.gondola.cc/tr:w-,h-,fo-auto/postThumbnails/57699aee6d.jpg
- **Posted:** 2026-09-18T14:42:12.000+00:00
- **Account Owner:** Aishwarya Srinivasan | Data & AI | LinkedIn Top Voice (@the.datascience.gal) — https://gondola.cc/the.datascience.gal

## Caption

AI’s Biggest Race Needs a Speed Limit

The CEO of Anthropic is asking AI companies to do the one thing no one wants to do in a race: slow down.

As AI starts helping build the next generation of AI, the real question is no longer just, “How capable is the model?” It’s: who checked that it is safe?

Dario Amodei’s most interesting proposal is to place independent evaluators inside AI labs, with ongoing access to the training process, internal tools, safety incidents, and the evidence behind company claims. And they must be free to publish what they find.

Because an audit means little if the company controls what auditors can see, or what the public gets to hear.

The hard part is getting every company, and eventually every country, to slow down together without someone secretly racing ahead.

So the measure of progress should not be days delayed. It should be access granted, findings disclosed, problems fixed, and whether the extra time produces measurably safer AI.

#AI #AISafety #Anthropic #TechPolicy #ArtificialIntelligence

## Stats

- **Views:** 24,590
- **Likes:** 221
- **Shares:** 0
- **Comments:** 4

## Tags

techpolicy, aisafety, anthropic, ai, artificialintelligence

---
Copyright (c) Gondola