[DigitalToday reporter Chi-gyu Hwang] Aidan Gomez (에이단 고메즈), CEO of enterprise AI company Cohere, warned that AI models are emerging as the "most powerful cyber weapon ever," able to find and exploit security vulnerabilities on a massive scale.
CNBC reported on Sept. 14 local time that Gomez said, "AI models are pushing the cybersecurity frontier into a new phase," adding that "AI models are extremely capable of finding and exploiting vulnerabilities."
Gomez presented using AI defensively as a solution. He said companies should use AI models to find and fix vulnerabilities in their own systems first. He stressed this is the top priority for protecting safety at present.
In some parts of the AI industry, calls are also growing to slow the pace of AI development. In a post released on Sept. 13, Anthropic CEO Dario Amodei argued that AI labs should slow the rate at which they improve model performance. He warned that while the direct damage from the OpenAI-Hugging Face incident was not large, it could have caused catastrophic harm if it had involved a group of misaligned agents with stronger capabilities.
Chi-gyu Hwang
delight@d-today.co.kr