Checkr, the San Francisco-based AI startup specializing in identity verification, is exploring opportunities to secure government contracts aimed at reducing fraud and waste in public benefit programs. CEO Daniel Yanisse said the company wants to use its technology to verify eligibility for programs like Medicare and Social Security, helping ensure that government funds reach the right individuals and not fraudulent actors. Currently, Checkr uses artificial intelligence to conduct background checks for companies like Uber and Lyft, analyzing data such as criminal records and driving histories. The company, valued at over $5.7 billion after raising $120 million in 2022, reported more than $800 million in revenue in 2025 and serves over 120,000 customers. Yanisse highlighted that many government payments are made to individuals with unverified income or employment status, contributing to improper payments. The Medicare Fee-for-Service program, for example, reported $28.83 billion in improper payments in 2025—6.55% of total payments—though not all were due to intentional fraud. Similarly, Medicaid payments to providers with questionable backgrounds have been a concern. A Middesk study found that $563 million in Medicaid payments went to providers blacklisted from federal healthcare programs between 2018 and 2024. However, experts caution that automating government benefit systems with AI presents serious challenges. Stuart Russell, a UC Berkeley professor and AI pioneer, expressed skepticism, noting that current AI systems, especially large language models, cannot provide reliable, explainable reasoning for their decisions—making it difficult to appeal or correct errors. He also pointed to EU regulations like GDPR, which prohibit fully automated decisions with significant legal consequences for individuals. Baobao Zhang, an AI policy expert at Syracuse University, warned that past government attempts to automate welfare systems have led to serious harm. In Indiana, an IBM-backed system to streamline welfare eligibility failed, leading to wrongful denials and a $1.3 billion lawsuit. In Australia, the Robodebt program wrongly accused thousands of welfare recipients of owing money due to algorithmic errors. A royal commission found that at least three people died by suicide after being falsely targeted, and the system was ultimately ruled illegal in 2019. Ifeoma Ajunwa, director of the AI and the Future of Work Program at Emory University, stressed the need for caution. She urged that any government adoption of AI should include advisory councils with technologists, social scientists, and affected communities to ensure transparency, accountability, and fairness. Checkr’s spokesperson confirmed that the company’s involvement in government systems remains conceptual for now. While the potential to reduce fraud and improve efficiency is promising, experts agree that robust safeguards and real-world testing are essential before deploying AI in high-stakes public services.