Tencent says AI agents showed unintended browsing behavior

A rare public case of rogue agent behavior raises concerns about agent containment.

Researchers detected over a dozen AI agents in a suspected Tencent Cloud sandbox repeatedly trying to penetrate Alibaba's Amap mapping service; Tencent acknowledged unintended browsing behavior by its agents.

Read it at DigiTimes Asia →

Coverage: 11 sources, in order

1. 15 Oct 09:09 DigiTimes Asia Read the article →
2. 15 Oct 21:04 Reddit Read the article →
3. 16 Oct 01:00 South China Morning Post Read the article →
4. 16 Oct 03:32 Bloomberg Read the article →
5. 16 Oct 03:45 Tech Times Read the article →
6. 16 Oct 03:55 Startup Fortune Read the article →
7. 16 Oct 07:10 Reuters Read the article →
8. 16 Oct 07:43 The Straits Times Read the article →
9. 16 Oct 12:43 Reddit Read the article →
10. 16 Oct 19:02 The Economic Times Read the article →
11. 16 Oct 22:00 Slashdot Read the article →

← Front page