Regeneron warns that use of agentic and autonomous AI systems may create operational, cybersecurity and data risks.
“We are increasingly utilizing AI-based solutions in various facets of our research, development, and other operations and continue to explore further use cases for AI. The use of AI solutions by our employees or third parties on which we rely may lead to the impermissible use or public disclosure of sensitive data. AI systems may generate inaccurate, incomplete, misleading, or biased outputs that may not be readily detectable by users, or may produce unreliable outputs over time. Reliance on such outputs could result in errors, inefficiencies, compliance issues, or other adverse consequences. Employee use of unapproved AI tools could create ungoverned data flows, unmonitored compliance gaps, and data protection risks that are not visible until an adverse event occurs. In addition, the increasing use of agentic and autonomous AI systems may create additional operational, cybersecurity, compliance, and data protection risks. Failures in the design, governance, or monitoring of such systems could result in unauthorized actions, inaccurate and/or biased outputs, disruption of business processes, or unauthorized access to sensitive information.”