UnitedHealth tracks employees' AI use

UnitedHealth Group has begun tracking how often some employees use artificial intelligence tools as the company pushes to embed the technology more deeply across its operations, according to a Bloomberg report published May 15. The report said some workers in Optum, the company’s health services division, are being monitored on whether they make at least one query a day in tools including OpenAI’s ChatGPT and Microsoft’s Copilot. UnitedHealth did not publicly announce the employee metric, but the company has separately described AI as part of a broader modernization effort. In first-quarter results released April 21, Chief Executive Stephen Hemsley said UnitedHealth was making “substantial artificial intelligence and cybersecurity investments.” ### Which employees are being measured, and on what standard? Bloomberg reported that some Optum workers are being tracked on whether they perform at least one AI query a day, citing people familiar with the matter. The report said the effort is part of a push to make AI a routine part of work across the company’s operations. (bloomberg.com) Optum is the unit that houses UnitedHealth’s pharmacy, care delivery, analytics and technology operations. UnitedHealth has not, in the public materials reviewed, disclosed how many employees are covered by the tracking, whether the measure applies companywide, or whether it is tied to performance reviews. ### How does UnitedHealth say it governs AI inside the company? (bloomberg.com) UnitedHealth says on its website that it uses a Responsible AI program with “processes, governance and monitoring” across AI systems. The company says it monitors AI use across the enterprise to understand how its tools are used and how they operate, and says its oversight includes clinical, legal, compliance, business and technical review. (sec.gov) The company also says it has an AI Review Board made up of clinicians, data scientists, business leaders, privacy and security experts, and ethicists. UnitedHealth says the board reviews AI models for safety, accuracy and fairness, while a proprietary platform called United AI Studio is used to build, deploy and manage applications. ### What does UnitedHealth require from outside AI vendors? (unitedhealthgroup.com) A UnitedHealth document labeled “Responsible Use of Artificial Intelligence – Minimum Standards” says vendors must maintain a governance program, conduct risk assessments and testing at least once each calendar year or after material updates, and carry out ongoing monitoring after deployment. The document also says vendors must keep records on validation, testing and ongoing operation. (unitedhealthgroup.com) The same standards require annual training for vendor personnel who support or use AI solutions, and say outputs must be validated for accuracy, completeness and bias, with human review before delivery to the customer. The document cites ISO 42001:2023, the National Institute of Standards and Technology AI Risk Management Framework and OECD AI principles. (unitedhealthgroup.com) ### Why is this happening during a period of broader scrutiny? UnitedHealth’s AI push comes as the company remains under regulatory scrutiny on several fronts. The Justice Department approved a settlement in December 2025 requiring divestitures to resolve its challenge to UnitedHealth’s $3.3 billion acquisition of Amedisys, and UnitedHealth said in July 2025 that it had begun complying with formal criminal and civil requests from the Department of Justice related to aspects of its Medicare business. (unitedhealthgroup.com) February 2024 is when the Justice Department opened an antitrust investigation into UnitedHealth’s relationship with Optum and physician-group acquisitions, according to Becker’s, which cited the probe in a 2025 report. UnitedHealth has not tied the employee AI tracking effort publicly to those investigations, but it has said its AI practices are designed to align with applicable laws, regulations and policy requirements. (justice.gov) ### Where has UnitedHealth said AI is already being used? UnitedHealth says AI is already being used in personalized search, customer messaging tools, call-center support and disease prediction. On March 26, UnitedHealthcare said an AI tool called Avery helps customer advocates answer questions on coverage, appointments, cost estimates, claims and benefits, and that the tool operates within governance frameworks with human oversight. (beckersasc.com) January 29 is when UnitedHealthcare said an AI-enabled program called Benefit Assist had delivered nearly four times more cash benefits to eligible members than other plans, according to the company. Those announcements show UnitedHealth discussing AI not only as an internal productivity tool but also as part of member-facing products and service operations. (unitedhealthgroup.com) April 21 is the latest date on which UnitedHealth gave investors a formal update that included AI spending. The company said then that it expected full-year 2026 adjusted net earnings of more than $18.25 per share and that it was continuing initiatives begun in the second half of 2025, including modernization and AI investment; its next quarterly update is expected with second-quarter results. (sec.gov) (unitedhealthgroup.com)