Healthcare AI Supply Chain Risks Spotlighted by MIT Study and EU Guidelines

A study by the MIT AI Risk Initiative surveyed international experts and identified AI‑enabled cyberattacks, dangerous capabilities and misinformation as top risks over the next five years, with the information and finance sectors seen as especially vulnerable. The report highlights how AI amplifies existing cybersecurity threats and creates new challenges for identity, data and compute perimeters.

In parallel, a breach at British healthcare‑software supplier Craneware exposed data from more than 2,000 health organisations and nearly 10,000 clinics, underscoring real‑world supply‑chain weaknesses in the sector. Federal U.S. agencies such as CISA and the FBI are shifting their defence posture from traditional perimeter security to focus on identity‑based safeguards and continuous monitoring.

European cybersecurity regulators are now urging health systems to close gaps in vendor contracts, demanding clear provisions for the handling and deletion of patient data, model embeddings and fine‑tuning datasets after a contract ends. The new procurement guidelines aim to prevent lingering proprietary AI artifacts that could compromise privacy or security when third‑party vendors depart.