IT Home On October 24, regarding cyber security issues in the AI era, billionaire, entrepreneur, and investor Mark Cuban proposed a rather interesting “elimination of threats” approach.
According to Business Insider’s report today (the 24th), Kuban compared the network attacks launched by large AI agents to cockroaches infiltrating a house: even if the house has just been sprayed with insecticide, cockroaches can still find a way to get in. “I often ask cybersecurity experts, and when hearing about various hacking attacks, technological breakthroughs, and security issues, I wonder whether AI has invented any new tactics? The answer is no. AI has not created new attack methods; it can only launch massive attacks, which are essentially brute force attempts.”
In Kuban's view, preventing AI network attacks is similar to preventing cockroaches from entering a house. “The principle is the same as preventing cockroaches. As long as someone creates enough AI agents, there will always be a way to find the system’s vulnerabilities and exploit them. The good news is that we can also ‘spray’ the system with a large amount of AI. I believe more people will invest effort in blocking malicious attacks. I am not pessimistic at all about AI.”
The artificial intelligence industry has also started to form its own “pest control teams” to prevent security risks. Over the past year, OpenAI’s number of security positions posted on LinkedIn increased by 91%, covering areas such as system testing, technical research, and governance.
According to IT Home, some risks have been revealed during testing. Anthropic disclosed that due to a configuration error in the third-party testing environment, three Claude models accessed the actual operating systems of relevant institutions without authorization. Since then, Anthropic has introduced a real-time classification model to identify and prevent AI models from attempting to break through the testing environment.
Kuban believes that although AI has lowered the barrier to launching malicious attacks, defenders can also use AI to enhance network security defenses. “Bad things will definitely happen, but cyberattacks have never been new. If attackers can launch attacks through large-scale brute-force attacks, defenders can respond in the same way.”