The advancement of AI in medicine: regulatory challenges in the digital age

The portal Medicina S/A published an article by Dr. Matheus Puppe, partner at Maneira Advogados, which discussed the advancement of the use of artificial intelligence tools by the medical and pharmaceutical sector.

The advancement of AI in medicine: regulatory challenges in the digital age

07/10/2023

By Matheus Puppe

The digital revolution has reshaped several sectors – and medicine and the pharmaceutical industry are not immune to its consequences. Recently, the world witnessed a remarkable advancement: the first medicine generated entirely by Artificial Intelligence (AI), which advanced to Phase 2 of clinical trials. This achievement, promoted by the pharmaceutical company, opened a new chapter in the history of medicine and reinforces the relevance of AI in the research and development of new medicines, with disruptive potential for the healthcare sector and capable of saving many people.

AI has enabled the analysis of massive volumes of data and, through complex algorithms, has been able to discover new targets for diseases and create novel molecules to act on these targets – something that would take decades for human scientists to achieve.

Now, this AI-imagined medicine will undergo the same rigorous clinical trials to ensure its efficacy and safety, in compliance with existing protocols.

Another development was the case in which an AI for commercial use was able to analyze various information from a test to arrive at the diagnosis of a rare disease, which went unnoticed by doctors.

However, the entry of AI into medicine does not come without challenges, especially in the regulatory field. The Food and Drug Administration (FDA) in the USA and the National Health Surveillance Agency (Anvisa) in Brazil, adopt a risk-based approach to the approval of medicines (risk based approach), that is, the benefits and risks of a medicine are carefully evaluated before its approval, however, the legislation is still outdated.

With AI-generated medicines, the approach is the same, but the complexity can be increased due to their unique and innovative nature.

In parallel, we see that the European Union's new AI regulation also categorizes AI systems into acceptable, limited, high and unacceptable risks. A similar approach to Brazil's AI regulation proposal, where both emphasize the importance of compliance with standards for the ethical use of AI.

Regulations whose specificities must inevitably be applied in the approval of medicines generated by AI, but which, due to their symmetry regarding the issue of regulated self-regulation, and a “risk based” approach, can have an interesting dialogue.

However, with the new developments in AI and technology, it is imperative to adapt and reformulate Brazilian legislation to accommodate the new changes in society, especially Anvisa regulations, which affect not only new medicines, such as these, but also medical software that uses AI, as some software is classified as Software as a medical device.

As an example, in a recent case from our office, a group of doctors created software based on machine learning systems and models, capable of emulating human intelligence, with the potential to assist doctors in a specific field around the world and maximize the treatment of different patients, however, it faced several regulatory obstacles from Anvisa, due to simple regulatory gaps. This case illustrates the urgent need to update the legal framework.

If, for example, software falls into the “Software as a Medical Device” category, this requires legislation adapted to new technologies. Given this scenario, it is worth noting that Brazil has the potential to lead these changes in the world and leave the passive pole in the race to regulate new technologies.

An example, although not dedicated to the life sciences sector, is the European Union's new AI regulation, and Brazil's AI regulation proposals. These are cases of efforts to regulate AI based on risks, but specific regulations such as those from Anvisa are left out at this time.

The fact is that AI is revolutionizing medicine, but current clinical and regulatory testing protocols, whether for medical devices, software or medicines, are not yet prepared for the new era of technology.

Therefore, it is essential that Brazilian legislation, in particular Anvisa regulations, is adapted and reformulated to accommodate these changes. Technological evolution demands an equally rapid response in the regulatory field. After all, the future of medicine has already arrived and we cannot let regulatory barriers prevent its full development, to the benefit of thousands of patients.

*Matheus Puppe, lawyer specializing in new technologies and partner at Maneira Advogados.