This week marked the start of a new era in cyber risk with threats of artificial intelligence hacking — and Senate Minority Leader Chuck Schumer fears that Washington isn’t ready.

Anthropic’s new frontier model, Mythos, cracked MacOS during testing, according to The Wall Street Journal. On Monday, Google’s Threat Intelligence Group said it thwarted the first known case of cybercriminals using AI to discover and target a previously unknown software vulnerability that could have allowed hackers to bypass two-factor authentication.

“There’s a discussion now that says, ‘it’s coming,’ right?” Google Threat Intelligence Group Chief Analyst John Hultquist said in an interview with Semafor. “That’s a misinterpretation of facts. The new era is here.”

Google engineers said they detected hackers targeting a “zero-day vulnerability,” a type of security weakness unknown to the software’s developers, which therefore leaves defenders “zero days” to deploy patches.

Schumer wrote to Homeland Security Secretary Markwayne Mullin earlier this month warning that new AI systems could dramatically accelerate cyberattacks against critical infrastructure, while criticizing the agency for its response thus far. He said he wants DHS to provide Congress with a national coordination plan for responding to “frontier AI-enabled hacking” by the start of July.