U.S. SENIOR OFFICIALS WARNED U.S. COMPUTER SYSTEMS ABOUT THE NEED FOR ENHANCED CYBER-DEFENSE TO RESPOND TO AI RISKS

On 6 August, the highest-ranking information security official of the United States Government warned that computer systems throughout the country needed to be strengthened to respond to the growing threat of AI-driven cyberattacks. In his speech on Wednesday, the Acting United States Chief Information Security Officer, Michael Duffy, stated that government and private sector institutions had in the past generally responded passively to cybersecurity. He stated that this approach had to be changed in order to take precautions before hacker attacks occurred in the AI era. “In view of these AI capabilities that we now see, we probably have no time to clean up the mess,” Duffy said in a thematic discussion at a conference on cybersecurity in Las Vegas, “The policy of the next decade cannot always follow up passively after a major event”. The rapid breakthrough of AI, including the Mythos model of Anthropic PBC, raised concerns that hackers could use this technology to launch destructive cyberattacks on critical infrastructure. The Trump Government has committed itself to taking action to address the network risks posed by AI and has recently completed a voluntary safety testing framework for front-line models。