The US government's most senior information security official warned that computer systems across the country need stronger defenses to deal with the growing threat of artificial intelligence-driven cyber attacks. Acting US Chief Information Security Officer Michael Duffy said in a speech on Wednesday that in the past, government and private sector agencies usually took a passive approach to cybersecurity. He said that to take preventive measures before hacker attacks occur in the AI era, this approach must be changed. “Given the AI capabilities we've seen so far, we probably don't have time to clean up the mess at all,” Duffy said during a panel discussion at a cybersecurity conference in Las Vegas. “Policies for the next ten years cannot always be passively followed up after a major event has occurred.” Including Anthropic PBC's Mythos model, rapid breakthroughs in AI systems have raised concerns that hackers may use this technology to launch destructive cyber attacks on critical infrastructure. The Trump administration is committed to taking action to address cyber risks posed by AI, and has recently completed a voluntary security testing framework for cutting-edge models.

The US government's most senior information security official warned that computer systems across the country need stronger defenses to deal with the growing threat of artificial intelligence-driven cyber attacks. Acting US Chief Information Security Officer Michael Duffy said in a speech on Wednesday that in the past, government and private sector agencies usually took a passive approach to cybersecurity. He said that to take preventive measures before hacker attacks occur in the AI era, this approach must be changed. “Given the AI capabilities we've seen so far, we probably don't have time to clean up the mess at all,” Duffy said during a panel discussion at a cybersecurity conference in Las Vegas. “Policies for the next ten years cannot always be passively followed up after a major event has occurred.” Including Anthropic PBC's Mythos model, rapid breakthroughs in AI systems have raised concerns that hackers may use this technology to launch destructive cyber attacks on critical infrastructure. The Trump administration is committed to taking action to address cyber risks posed by AI, and has recently completed a voluntary security testing framework for cutting-edge models.