Chinese AI tool told researchers how to make bioweapons
12 Articles
12 Articles
A SECURITY firm has discovered that a Chinese AI model called Kimi can bypass protection and give instructions for making biological weapons and assassinations. The maker of Moonshot has launched an internal investigation.
This was discovered during the "Jailbreaking" process, where researchers use a series of complex instructions to see if artificial intelligence tools ignore set restrictions.
According to a BBC report, Mindgard, a company that tests the security of artificial intelligence systems, found in its July studies that Moonshot's Kimi K2.6 and K3 Swarm models had some security flaws implemented by the developers...
The Chinese artificial intelligence developer Moonshot conducts an internal analysis after researchers have managed to convince two of his popular Kimi models to provide information about the manufacture of biological weapons, writes BBC.
Chinese AI developer Moonshot has launched an internal audit after security researchers managed to fool theirs.
Coverage Details
Bias Distribution
- 50% of the sources lean Left, 50% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
