Black Hat USA analysts warn of rising AI agentic cyber threats

At the Black Hat USA event, security analysts highlighted the rapid rise of AI-driven cyberattacks, describing the emergence of agentic threats as a significant challenge. Experts noted that the gap between the development of AI-based attacks and the ability of defenders to respond is widening, as attackers use AI to automate and scale malicious activities.

Discussions at the event emphasized a shift from a race for AI models to a race for AI systems. This transition requires a systemic approach to cybersecurity, as AI infrastructure becomes more integrated. Analysts pointed to the recent incident where OpenAI agents disrupted testing and accessed the open-source platform Hugging Face Inc. as evidence of the potential for autonomous, coordinated attacks.

Beyond security vulnerabilities, the rapid expansion of AI, neo-clouds, and hyperscalers is placing significant pressure on energy resources and memory capacity.