Advances in artificial intelligence could increase the scale and complexity of cybersecurity threats as autonomous agents make sophisticated attacks easier to replicate, according to analysts at Evercore.
The firm said sophisticated cyberattacks have historically been constrained by the amount of human labour required to identify targets, discover vulnerabilities, develop exploits and respond to defensive measures.
Evercore expects agentic AI to reduce some of those constraints by allowing advanced capabilities to be deployed across large numbers of autonomous or semi-autonomous agents.
AI Could Increase Volume of Cyberattacks
According to Evercore, AI systems could enable thousands of agents to continuously identify, probe and potentially exploit vulnerabilities, compared with individual attackers targeting a more limited number of systems.
The analysts said this development could temporarily shift the balance between cyberattackers and defenders towards attackers in the near term.
Evercore also expects the adoption of frontier AI models within businesses to expand the potential attack surface and increase the number of machine and agent identities requiring security controls.
The firm said these developments could increase both the potential speed and scale of cyberattacks.
Evercore Sees Continued Cybersecurity Spending
Evercore continues to view cybersecurity as an area of sustained enterprise technology spending as companies adapt their security infrastructure to AI-related risks.
The analysts identified identity security as one area that could see increased demand as businesses manage access across employees, autonomous agents and other non-human identities.
Companies with exposure to this area include Okta (NYSE:OKTA), Palo Alto Networks (NASDAQ:PANW), CyberArk (NYSE:CYBR), SailPoint (NYSE:SAIL) and CrowdStrike (NASDAQ:CRWD), according to Evercore.
The firm’s assessment reflects its expectations for the development and adoption of AI systems and their potential implications for enterprise cybersecurity.