AI as your new medical assistant: who pays the bill if something goes wrong?

The impact of AI in medicine is huge. As well as improving diagnoses, the technology also optimises care processes and treatments. Here are a few examples of what it offers:

• Faster, more accurate diagnoses: Algorithms analyse medical images such as CT scans or mammograms with impressive precision. They detect abnormalities that are difficult for the human eye to detect, leading to faster and better diagnoses.
• Predictive and personalised care: By analysing large volumes of patient data, the software learns to spot patterns that indicate an increased risk of diseases such as sepsis. This makes preventive interventions and personalised treatment plans possible, for example by predicting how patients will respond to chemotherapy.
• More efficient hospital management: The technology automates repetitive tasks such as planning and record management. It also helps predict bed and staff demand, allowing for more efficient use of hospital resources.
• Better patient support: Smart chatbots keep patients informed and enable remote monitoring – particularly useful for patients with chronic illnesses.

This technological advance creates a complex legal problem. Doctors used to take the decisions as the sole expert; now, though, there is a highly influential digital assistant. Several important questions are raised:

• Who bears ultimate responsibility? Do doctors always remain ultimately responsible, even if they rely on the output of a certified system?
• When is the product defective? Can the hospital hold the software developer liable if the algorithm doesn’t provide the promised security?
• What is expected of the doctor? Should doctors be prepared to ignore an algorithm’s recommendation? And what if this leads to harm?
• What’s the role of the hospital management? The management is responsible for purchasing secure technology. Directors’ and officers’ liability may be triggered by a bad choice.

These new risks have a direct impact on your hospital’s and doctors’ insurance.

1. Professional and commercial civil liability: Under Belgian law, AI can be considered ‘auxiliary equipment’. This means that the healthcare provider is in principle liable for its failure, just as with any other medical instrument. The ultimate responsibility therefore often remains with the doctor, who must interpret and validate the output.
2. Directors’ and officers’ liability: Directors are responsible for strategic investments. When purchasing AI systems, they must check their reliability and certification thoroughly. If they act negligently, they may be held personally liable for a management error.
3. Cyber ​​insurance: AI systems are vulnerable to cyber attacks. A cyber policy is essential for covering the direct damage arising from an incident, but it usually doesn’t cover the consequential damage of a medical error caused by a hacked system.

Don’t wait for the upcoming European AI Act. Take proactive steps now to manage the risks:

• Thorough screening before purchase: Ensure that suppliers and their products are thoroughly checked.
• Unambiguous contracts: Determine liability contractually with the software vendor. Who is responsible for updates and bug fixes?
• Clear protocols and training: Develop internal protocols for using the software and make sure your employees are given thorough training that covers both its possibilities and its limitations.
• Human oversight remains crucial: Emphasise that AI is a tool, not a replacement for clinical judgement. The doctor must always make the final decision.

The rise of AI in healthcare is a positive development. However, a proper understanding of the risks and a proactive approach are crucial to making the most of the opportunities while simultaneously ensuring patient safety and protecting your institution. As your insurance broker, we keep a close eye on these developments and proactively advise you on the best way to adapt your policies to the risks of tomorrow.