Artificial Intelligence Poses Threat to Patient Privacy

One concern is that AI could make it possible to re-identify patients from anonymized data. Shutter Stock

Specific regulations for artificial intelligence (AI) technologies may be warranted to ensure the confidentiality of patient data. Currently, hospitals may use anonymized patient data or share or sell it to further development of AI, which can pose risks to patient privacy. “AI also makes it easier to re-identify patients from de-identified data that’s been shared by triangulating from different data sources,” according to W. Nicholson Price II, a professor of law at the University of Michigan Law School in Ann Arbor, who specializes in intellectual property, health law, and regulation.

He pointed out that potential risks to patients may not have manifested yet, but could occur as the AI technologies become powerful. Businesses and consumers and others affected by these systems have a right to know if technologies have been adequately vetted and risks have been appropriately mitigated, according to Price.

Health care providers, Price said, need to be careful about how they share and use patient data. “The HIPAA safe harbor [rules] means that sharing de-identified patient data is still probably okay for now from a HIPAA perspective, though that could potentially change,” he said. “We need to have a conversation about how health data are used to develop AI. Right now, it’s too easy to get data for some entities and too hard for others. We’re not getting the balance right, and that needs to change. How we get to a better place, though, is a really tough question.”

Niam Yaraghi, PhD, an assistant professor of business technology at Miami Herbert Business School of the University of Miami in Florida, agrees, saying we are in a new era in terms of policing protected health information (PHI). In principle, AI is a set of statistical methods designed to uncover patterns in data. With recent advances in this area, it is now easier than ever to uncover patterns that we may have not been aware of in the past.

“This could create a major threat to re-identification of anonymized data, especially when such data are merged with other sources of data across multiple platforms,” said Dr Yaraghi, who also is a Senior Fellow at the Brookings Institution’s Center for Technology Innovation in Washington, DC. “Our privacy protection laws, including HIPAA, are outdated and do not have provisions to help protect the privacy of patients in these new situations.”

One of the biggest issues facing physicians is educating their employees and vendors. Many physicians may not realize how their patient data are being used, Dr Yaraghi said. “As AI makes it easier to mine the data, the benefits for the AI users increase, while the privacy risks for the patients also increase. It is a win-lose game between the data miners and the patients,” he said.

Physicians should not underestimate the threat AI technology poses to medicine and the delivery of care, he noted. “They have to be extra vigilant about the possibility of re-identification when it comes to sharing of their data outside their organization,” he said. “They can no longer assume that their older methods of masking the data would still protect the privacy of their patients.”

On April 25, 2023, 4 federal agencies jointly pledged to uphold America’s commitment to the core principles of fairness, equality, and justice as emerging automated systems and AI become increasingly common in peoples’ daily lives. It is impacting not only patient privacy but also civil rights, fair competition, consumer protection, and equal opportunity. The Civil Rights Division of the United States Department of Justice, the Consumer Financial Protection Bureau, the Federal Trade Commission, and the Equal Employment Opportunity Commission released a joint statement outlining a commitment to enforce their respective laws and regulations.

All 4 agencies have previously expressed concerns about potentially harmful uses of AI and resolved to vigorously enforce their collective authorities and to monitor the development and use of automated systems. Now that AI has spread to every corner of the economy, it is paramount that regulators stay ahead of its growth.

This article originally appeared on Renal and Urology News