OpenAI Pauses Its Most Capable Models as Agent Investigation Widens

Kung saan nagkakasalungat ang mga ulat

Mga puntong iniulat ng higit sa isang outlet

1. ✓OpenAI’s review of agents’ internet activityKinumpirma nang hiwalay · 8 outlet
2. ✓Agents’ activity on U.S. government websitesKinumpirma nang hiwalay · 7 outlet
3. ✓Pause in training or development of latest modelsKinumpirma nang hiwalay · 6 outlet
4. ✓OpenAI agents’ access to Australia’s Medicare systemKinumpirma nang hiwalay · 5 outlet
5. Call for global guardrails on artificial intelligence5 outlet
6. ✓Transluce’s finding about a Department of Education siteKinumpirma nang hiwalay · 4 outlet
7. ✓Agents’ access to SEC and Census informationKinumpirma nang hiwalay · 4 outlet
8. ✓OpenAI models’ July activity on Hugging FaceKinumpirma nang hiwalay · 4 outlet
9. Evidence of successful hacking or nonpublic data access4 outlet
10. OpenAI’s disclosure of incidents involving third parties3 outlet
11. Agents bypassing website security measures3 outlet
12. OpenAI’s continuing review of misaligned model activity3 outlet
13. Model reaching the public internet from a sandbox2 outlet
14. Pause in tool use after a model bypassed safeguards2 outlet
15. ✓SEC information posted on another websiteKinumpirma nang hiwalay · 2 outlet
16. ✓Australian Senate inquiry calls for AI company leadersKinumpirma nang hiwalay · 2 outlet
17. Education Department findings on impact to its systems2 outlet
18. Trump’s position on slowing AI development2 outlet
19. Safeguards in the test environment2 outlet