UnitedHealth tracks employee AI use

UnitedHealth Group is not just rolling out more AI tools. It is reportedly measuring whether some employees are using them every day. Bloomberg reported on May 15 that UnitedHealth is tracking how often some workers in its Optum services division use tools such as OpenAI’s ChatGPT and Microsoft’s Copilot, citing people familiar with the matter. The aim, according to the report, is to embed AI more deeply across the company’s operations. (bloomberg.com) That detail matters because it shifts the story from AI experimentation to workplace management. Usage is being treated as something the company can count, compare and push. Bloomberg said the company is monitoring whether some workers perform at least one AI query a day. (bloomberg.com) UnitedHealth has been publicly presenting AI as a broad operating priority. On its website, the company says it uses AI to simplify patient navigation, reduce administrative work for clinicians and support more timely insights. It also says it has a framework for responsible AI use in health care. (unitedhealthgroup.com) The timing is hard to separate from the company’s recent cyber history. UnitedHealth told the SEC on February 21, 2024, that it had identified a suspected nation-state-associated cyber threat actor in Change Healthcare systems and had isolated affected systems. That attack disrupted parts of U.S. healthcare payments and operations because Change handles a large volume of claims and transaction traffic. (sec.gov) In that context, tracking employee AI use can be read less as a novelty than as a control question: which tools are being used, by whom, and with what oversight. UnitedHealth has not publicly framed the reported tracking that way in the sources reviewed, but the company’s own AI materials emphasize responsible use, governance and deployment in sensitive healthcare settings. (bloomberg.com) Healthcare companies face a narrower margin for error than many other large employers. Staff may touch claims data, care-management workflows, customer service records and provider systems. Any push to normalize generative AI in those environments raises practical questions about audit trails, approved tools, prompt handling and whether workers are using consumer AI products or enterprise versions with tighter controls. That is an inference from the setting and the reported monitoring, not a stated company rationale. (bloomberg.com) UnitedHealth is also signaling that AI is already embedded in customer-facing products. In March, UnitedHealthcare introduced “Avery,” which it described as a generative AI companion for members and customer advocates. In January, the company promoted an AI-enabled program called Benefit Assist. Those announcements show the AI push is not limited to internal experimentation. (unitedhealthgroup.com) The broader healthcare security backdrop has also become more focused on disruption, not just data loss. A Black Book Research survey released May 16 said 82% of European hospital cybersecurity buyers reported very high or extreme concern about attacks, and the study said hospital cyber risk is increasingly being evaluated as a clinical continuity issue. The survey is Europe-focused and not a direct measure of UnitedHealth, but it underscores the operating environment in which large health companies are making technology-control decisions. (newswire.com) UnitedHealth’s next scheduled investor checkpoint is its regular reporting cycle after the company issued its 2026 outlook on January 27. What is public so far shows a company trying to scale AI in a heavily regulated business while keeping closer count of how employees actually use the tools. (unitedhealthgroup.com)