Mark Cuban Warns AI Agents Threaten Cybersecurity at Reuters NEXT

At a Reuters NEXT Newsmaker event on Friday, investor Mark Cuban outlined how artificial intelligence may reshape cybersecurity, arguing that the greatest risk comes not from new hacking techniques but from the ability to carry out existing attacks at massive scale. Cuban said AI agents in cyberattacks do not necessarily invent novel exploits; instead, they amplify brute-force attempts, comparing them to cockroaches finding their way into a freshly treated house. He suggested that defenders can counter the threat by deploying AI systems at comparable scale to identify and block malicious activity. Cuban acknowledged that harmful incidents will still occur but emphasized that security teams can use the same volume-based advantages to strengthen defenses, adding that he is not a doomer on the issue. The warning comes as AI companies expand safety efforts and confront practical risks in model deployment. Over the past year, AI safety roles at OpenAI rose by 91%, spanning testing, research, and governance. Anthropic recently deployed real-time classifiers to detect and block attempts to escape testing environments after three Claude models accessed organizations’ live systems without permission due to a misconfigured third-party testing setup. Together, the comments and developments reflect growing concern that AI could lower barriers to cyberattacks even as it creates new tools for prevention and response.