Jamie Dimon, CEO of JPMorgan Chase, said Anthropic’s Mythos AI model significantly increases cybersecurity risks.
Dimon told Bloomberg Television on Tuesday that the risks “increased tenfold after the myth.”
“Artificial intelligence creates vulnerabilities that we don’t know about, and before these things we were always worried about networks,” he added.
The JPMorgan CEO has long described cybersecurity as one of the bank’s biggest risks, and his latest comments stand out as the technology industry grapples with the impact of advanced artificial intelligence models.
Dimon is the latest prominent business leader to warn of the risks posed by increasingly powerful artificial intelligence systems.
His comments come as cutting-edge artificial intelligence developers face intense scrutiny over the safety of their most advanced models.
In recent weeks, artificial intelligence researchers, technology executives and some lawmakers have renewed calls for greater international cooperation to reduce the risk of catastrophic misuse of advanced artificial intelligence.
Anthropic and OpenAI revealed separate incidents that occurred during security testing in recent months, raising new questions: How reliable are the controls of their most powerful artificial intelligence models?
The incidents have also prompted some AI researchers and executives to call for slowing the development of cutting-edge AI models until stronger security measures are put in place.
Dimon said that these risks are real, but he believes that practical safeguards should be adopted to deal with these risks, rather than panic.
“The downside is obviously what you read about agents and mythology and all these things that could cause trouble, it’s a legitimate concern, it’s a real thing,” he said.
“I’m not hysterically asking, ‘Does this exist?’ What we’re doing is rolling up our sleeves and trying to figure this out,” he said.