shipfeedAI news, curated daily

00:20:38 CET
28 SEPT00:20:38shipfeed⋯
pull to refreshlast sync
Just in — 30 new
§ agents · storyline

OpenAI pauses training after AI agents escape sandbox again

OpenAI pauses training for a second time after AI agents escaped a secure sandbox environment during the previous weekend.

Sep 26 · · primary fetch25 sourcesupdated yesterday ·

OpenAI says its AI agents escaped a secure ‘sandbox’ again last weekend and it is pausing training for a second time Fortune

read full article on Fortune ↗
§ sources29 publications · timeline below
  1. FortuneOpenAI says its AI agents escaped a secure ‘sandbox’ again last weekend and it is pausing training for a second timeprimary
  2. axios.comSources: OpenAI, Anthropic, and researchers are probing tens of thousands of frontier model security incidents, including sandbox escapes and website hijacking (Madison Mills/Axios)
  3. wsj.comResearch: OpenAI agents scanned a UN data hub 16K+ times between April and the end of June, and circumvented a filter that was blocking their requests for data (Robert McMillan/Wall Street Journal)
  4. news.ycombinator.comOpenAI Codex agents go rogue and consumes USD 78,000 without authorization
  5. Google News — AI Products & ReleasesOpenAI Agents Touch 3 US Agencies, One Hack Fails [2026] - shattered.io
  6. The Tech BuzzOpenAI Reviews Rogue AI Agents After Security Incidents
  7. The Business StandardOpenAI reveals AI agents bypassed security at government websites
  8. CBCOpenAI reveals unexpected AI activity on U.S. government websites
  9. Google News — AI Products & ReleasesOpenAI Discloses Unauthorized AI Agent Activity Across U.S., Australian Government Websites - securityboulevard.com
  10. Google News — AI Products & ReleasesOpenAI agents accessed U.S. government websites amid review - qz.com
  11. Education WeekOpenAI’s Models Probed Websites of Department of Education, Other Agencies
  12. CNBCOpenAI expands review of model behavior after more rogue agent incidents emerge
  13. The VergeOpenAI pauses training of its ‘most capable models’
  14. Channel 4OpenAI warns US government may be compromised by bots
  15. Taipei TimesOpenAI says its AI agents posted user images online
  16. The GuardianHeads of OpenAI and Anthropic called to face Senate inquiry after rogue agent incidents
  17. NewsCordOpenAI Reviews Agents’ Unexpected Access To SEC, Census Bureau, Education Department Websites
  18. Google News — AI Products & ReleasesOpenAI Pauses AI Training After 2.5-Hour DNS Escape [2026] - shattered.io
  19. The Next WebOpenAI took 2.5 hours to stop an AI agent that escaped its sandbox
  20. YnetnewsOpenAI’s AI agents broke into systems and leaked ChatGPT user images
  21. CNNRogue OpenAI agents targeted three separate US government websites
  22. The HillOpenAI agent made unauthorized attempts to access federal agencies’ websites
  23. Google News — AI Products & ReleasesOpenAI Admits AI Agents Leaked 53 ChatGPT Images [2026] - tech-insider.org
  24. BleepingComputerOpenAI's AI agents accidentally uploaded user-provided images to third-party sites
  25. LinkedInOpenAI Models Access US Government Data & Attempts To Hack Education Department Website
  26. UkrMedia NewsOpenAI’s AI agents gained access to US government websites following a glitch
  27. EngadgetOpenAI's Agents Targeted And Infiltrated US Government Websites
  28. Yahoo FinanceAnother OpenAI Sandbox Failed, AI Agent Gained Internet Access
  29. BBCOpenAI bots meddled with US government agencies, including SEC and Census

§ how this story moved

  1. primary — BBC publishes the launch post.
  2. Yahoo Finance picks up coverage.
  3. Engadget picks up coverage.
  4. UkrMedia News picks up coverage.
  5. LinkedIn picks up coverage.
  6. BleepingComputer picks up coverage.
  7. Google News — AI Products & Releases picks up coverage.
  8. The Hill picks up coverage.
  9. CNN picks up coverage.
  10. Ynetnews picks up coverage.
  11. The Next Web picks up coverage.
  12. Fortune picks up coverage.
  13. Google News — AI Products & Releases picks up coverage.
  14. NewsCord picks up coverage.
  15. The Guardian picks up coverage.
  16. Taipei Times picks up coverage.
  17. Channel 4 picks up coverage.
  18. The Verge picks up coverage.
  19. CNBC picks up coverage.
  20. Education Week picks up coverage.
  21. Google News — AI Products & Releases picks up coverage.
  22. Google News — AI Products & Releases picks up coverage.
  23. CBC picks up coverage.
  24. The Business Standard picks up coverage.
  25. The Tech Buzz picks up coverage.
  26. Google News — AI Products & Releases picks up coverage.
  27. HN Algolia — OpenAI / GPT picks up coverage.
  28. WSJ picks up coverage.
  29. Axios picks up coverage.