The advancements of technology and artificial intelligence in particular have led to the promise of better disease prediction through algorithms, more reliable surgery methods through robots and more widely accessible healthcare thanks to artificial intelligence. This leads regulators in the industry to question the legal risks of digitalisation in one of the largest and most important areas of work in society.

The potential for artificial intelligence in the field of health is immense, but it leads us to ask, “Who is accountable if the self-learning algorithm makes a mistake?” amongst many other questions such as origin of data protection and regulatory classification in the grey areas of health digitalisation.

The prospect of using artificial intelligence in health is one that sounds particularly appealing at first: It works at an incredible pace to detect early symptoms and perfect diagnostics. This pace of development is allowing AI to complete the same tasks as humans, but at a faster rate and a lower cost. AI also provides customers with a better and more accessible control of their health and wellbeing – an example being fitness tracker apps and digital accessories. Pattern recognition can identify patients with a high risk of specific diseases at incredibly fast rates through predictive analytics and analysis of the patient’s lifestyle, environment and DNA. On the other hand, AI is aiming to assist with monitoring treatment successes and conducting difficult surgeries through robot technology.

While all of these uses sound incredibly attractive to both doctors and patients alike, from a legal perspective, their development would mean entering into unknown legal territory. First of all, it would be necessary to clarify the regulatory classification of the product – is the software a medical device? Since medical devices can only be marketed if they carry a CE label after having undergone a conformity assessment procedure, it is important to legally clarify the classification. If the product does not have CE labelling, it could be under threat of competitors demanding it be pulled off the market or may even have criminal consequences due to administrative breaches. According to the German Medical Devices Act, if the software is intended to treat or detect diseases, it should be classified as a medical device. However, if the device only provides knowledge or stores data it will not qualify as a medical device.

Data protection is another relevant issue in the digitalisation of the health industry. AI solutions, specifically deep learning applications, are generally established through the continuous and in-depth analysation of specific patient medical data. Since May 2018, the requirements for collection of personal data have increased due to the European General Data Protection Regulation. This makes patient data a lot more difficult to collect and use in data analysis.

Furthermore, in civil law, there is a large area that is still to be clarified when it comes to AI and robotics based on AI. For example, the liability for errors and injuries, regulation and validity of contracts and insurance.

On the 16th of February 2017, the European Parliament adopted a resolution with recommendations to the Commission on Civil Law Rules and Robotics which will aim to clarify the question of with whom a treatment contract is made when a robot is used during surgery, who is liable if an AI commits and error which causes harm to the patient, and how patient data is shared to optimise the use of the AI software in a patient’s medical procedures.

The most difficult part of regulating algorithms is the deep learning ability it has. Establishing approval for an algorithm is difficult enough, but once that algorithm updates and learns from more data, this algorithm adapts and is no longer the same algorithm, therefore requiring a whole new regulatory procedure. The new learning cannot be implemented in the clinical practice without updated approvals.

Regardless of the legal unclarity in the utilisation of AI in healthcare, the potential is indisputable. It will aim to provide faster, cheaper and more efficient access to healthcare for people all over the world. The unstoppable progress of digitalisation may be the solution we all need towards a healthier world population, that is, if the law manages to keep up.

Ai could follow prescriptions and our internal conditions at all times

we already rely on
technology for the
monitoring of our

everyday health

Ellie Nikolova

Digitalisation in Healthcare – Artificial Intelligence and the Legal Risks Around It