Rogue AI saga continues: Chinese AI model Kimi K3 reportedly escapes during security test
A security test of Moonshot AI's Kimi K3 has uncovered another case of an advanced AI model operating beyond its intended limits. Researchers say the incident, enabled by a sandbox misconfiguration, highlights both the growing capabilities of AI agents and the importance of robust safeguards around their deployment.
Related Articles
0Centre asks Meta to curb deepfakes, examines role of recommendation algorithms
The Centre has stepped up scrutiny of Meta’s operations in India, asking the company to take stronger measures against deepfakes and other harmful art...
0Why Coimbatore is planting bamboo around its lakes
Can bamboo help save our lakes? Around Vellalore and Kolarampathy lakes in Coimbatore, 16 species of bamboo are taking root acting as a natural green...
010 tablets that can replace your laptop: Top Amazon Great Freedom Sale picks with attractive discounts and offers
Upgrade your work setup with feature-rich tablets offering laptop-like performance, now available at attractive prices during the Amazon Great Freedom...
0Towards nutritional literacy: navigating food labels, social media feeds and traditional wisdom to arrive at better diets
As consumers, we need to move beyond simplified narratives that classify every food as either ‘clean’ or ‘toxic’. Instead, we should understand what p...