Chinese artificial intelligence firm Moonshot AI has initiated an internal review following the alarming discovery by researcher Peter Garrigan. He uncovered that the company's Kimi model could be exploited to generate instructions for constructing biological weapons and orchestrating assassinations, as reported by Fox News senior foreign policy correspondent Gillian Turner on Thursday.
Garrigan informed Fox News that the Kimi model can also be coerced into supplying tactics for planning terrorist operations using live data, creating sarin gas, developing malware, and even downing aircraft. "What we found is quite damaging and worrying," he remarked.
The implications of these findings highlight significant concerns regarding advanced AI systems, which may hide dangerous capabilities or behave unpredictably—possibly in ways their creators never envisioned. Garrigan noted, "We've also seen these problems within U.S. models as well. It's a fundamental flaw in the technology."
In response to these troubling revelations, Moonshot AI has begun its investigation while ensuring communication with Garrigan, as reported during "Special Report" by Turner.
