Lawyers
Global Artificial Intelligence Group Co-Chair Katherine Forrest spoke with Lexology Pro about the heightened legal and security risks of AI agents. In “AI risks associated with OpenClaw: key lessons for businesses,” published on February 20, Katherine says that tools such as the open-source AI agent OpenClaw could expose users to cyberattacks and data breaches. “Most agentic tools actually access other datasets outside of a company, what I call ‘opening the door to the organization,’” Katherine notes. These AI tools can “walk outside to search the web or gather information from other sources and then return, which raises entirely new cybersecurity risks.”
She adds that, when organizations are hesitant to deploy agentic AI tools due to these risks, “shadow AI” may emerge, with employees seeking out personal alternatives to enterprise-grade tools. Companies can combat this by “giving employees enterprise versions of the tools they want,” she notes.
Related Insights
