Your Medical Scans Could Fuel AI – Here’s How to Protect Your Privacy
Artificial intelligence is becoming a standard tool in radiology. Many hospitals now use AI to help radiologists read X-rays, MRIs, and CT scans. The technology can speed up diagnoses and catch subtle findings. But it also creates new privacy risks for patients—ones that most people are unaware of. A recent warning from the Radiological Society of North America (RSNA) highlights why this matters now.
What happened
In May 2026, the RSNA published an article titled “Medical Imaging AI Opens a Pandora’s Box of Privacy-Related Risks.” The piece details how AI systems that analyze medical images can inadvertently expose sensitive health information beyond what is visible in the scan itself. For example, body shape, bone structure, and even facial features can be reconstructed from imaging data. That data can then be combined with other records to re-identify patients, even if their names are removed.
The article also points out that many AI tools are developed or hosted by third-party vendors. Images may be sent to cloud servers for processing, sometimes without the patient’s explicit consent. And those images can be used to train future AI models, meaning your scan could end up in a training dataset for a commercial product you never agreed to.
The RSNA warning is significant because it comes from the leading professional society for radiologists. It reflects growing concern within the medical community, not just from privacy advocates.
Why it matters
For most people, a medical scan is a routine, private event. You expect the images to stay within your healthcare provider’s network. But AI changes that expectation in a few ways.
Data breaches. When images are uploaded to cloud platforms for AI analysis, they become part of a larger digital system. Any system with network access can be breached. In 2024, a major health AI vendor suffered a data leak that exposed millions of patient scans. The risk is not theoretical.
Re-identification. Even when personal identifiers are stripped, researchers have shown they can re-identify patients using facial features visible in head CTs or even the unique geometry of a person’s spine. Once identified, that data can be linked to insurance records, employers, or other databases.
Insurance and employment discrimination. If an insurer learns that your scan was analyzed by an AI that flagged a possible (but not definite) early sign of a condition, it could affect your premiums or coverage. Current laws like HIPAA have gaps when data is used for AI training rather than direct treatment.
The RSNA warning underscores that these risks are not hypothetical. As AI use in radiology expands, the privacy controls have not kept pace.
What readers can do
You don’t have to avoid medical scans. But you can take steps to protect your data.
Ask who is using AI on your images. Before a scan, ask your doctor or the imaging center: “Are you using AI to analyze these images? Who provides the AI software? Where does my image data go?” You have a right to a clear answer.
Opt out of research and non-clinical use. Many consent forms include a clause that allows your data to be used for research or product development. You can ask to sign a version that restricts use to your own care only. Not all facilities offer this, but it’s worth requesting.
Use facilities with strong privacy policies. Large academic medical centers often publish data handling policies online. Look for statements about AI use, data retention, and third-party vendors. If you can, choose a provider that processes AI on-site (locally) rather than in the cloud.
Monitor your medical records. After a scan, request a copy of your imaging report and check that no additional AI analysis has been added without explanation. If something seems off, ask your provider.
Push for stronger regulations. Contact your elected representatives and support organizations like the Electronic Frontier Foundation or Patient Privacy Rights that advocate for health data protections. The RSNA warning is a wake-up call for lawmakers too.
Sources
- Radiological Society of North America. “Medical Imaging AI Opens a Pandora’s Box of Privacy-Related Risks.” May 20, 2026.
- RSNA news coverage on Google News, May 20, 2026.
Note: Some specific details about AI vendor data leaks and re-identification techniques are drawn from publicly reported research and news sources; individual facilities’ practices may vary.