Active mapping of artificial intelligence agents' critical systems access has been conducted by only 30% of U.S. businesses, which could increase the risk of insider security threats, TechRepublic reports.
Healthcare, which has accelerated the adoption of AI agents for protected health information and medical systems management, is particularly at risk, with 61% reporting identity-related intrusions and 42% disclosing a failed identity-related compliance audit, according to a BeyondID survey. "AI agents don't need to be malicious to be dangerous. Left unchecked, they can become shadow users with far-reaching access and no accountability," said the report. Such findings come as agentic AI has been increasingly implemented by organizations following a surge predicted by OpenAI CEO Sam Altman. "We may see the first AI agents 'join the workforce' and materially change the output of companies," Altman said. Meanwhile, integration of agentic AI in enterprise software apps has been expected by Gartner to spike from less than 1% last year to 33% by 2028.
Healthcare, which has accelerated the adoption of AI agents for protected health information and medical systems management, is particularly at risk, with 61% reporting identity-related intrusions and 42% disclosing a failed identity-related compliance audit, according to a BeyondID survey. "AI agents don't need to be malicious to be dangerous. Left unchecked, they can become shadow users with far-reaching access and no accountability," said the report. Such findings come as agentic AI has been increasingly implemented by organizations following a surge predicted by OpenAI CEO Sam Altman. "We may see the first AI agents 'join the workforce' and materially change the output of companies," Altman said. Meanwhile, integration of agentic AI in enterprise software apps has been expected by Gartner to spike from less than 1% last year to 33% by 2028.
