AI in MedTech

KI/AI in MedTech

With the launch of ChatGPT on November 30, 2022, artificial intelligence (AI) experienced a huge boom worldwide. Within two months, 100 million people had registered for the chatbot developed by OpenAI. The already prominent topic developed into a megatrend in a short space of time. ChatGPT and competing products such as Google’s Gemini are generative AIs. They create new content like text, images, and music by learning from patterns of existing data. Part of the fascination with AIs is their impact on the workplace.
AI also has the potential to have a major impact on the MedTech industry. According to an article by the Boston Consulting Group, the use of artificial intelligence is increasing in no other sector as much as in healthcare. The possibilities offered are as extensive as artificial intelligence is complex. But what areas of application are there for AI in the MedTech sector? What disadvantages can arise? And what regulatory requirements need to be observed? This article aims to explore these questions.

Application of AI in the MedTech industry

Areas of work in most companies receive optimization with the help of AI. In software development, codes can be generated, checked, and improved. The marketing & sales department can use artificial intelligence to conduct competitive analyses or create media and text documents. In customer support, it is also realistic that AI will take on far more tasks in the future than it does now. With access to relevant information and machine learning, artificial intelligence can answer complex customer queries independently in the future. Many departments could benefit considerably from a reduction in workload. However, these approaches are not exclusive to this industry.

Generative product design is a different matter. The idea behind it is also not exclusive to the MedTech industry, but the application is customizable. For example, new ideas for products and functions are developed based on existing patterns. Among other things, this creates unconventional solutions that serve as a source of inspiration for designers. Subsequently, the MedTech sector will see personalized products created with AI. This potential is already being used to manufacture bone implants. Another conceivable area of application is proactive post-market surveillance. Artificial intelligence has the potential to carry out literature research and the subsequent creation of reports faster and with fewer errors.

As AI development continues, the number of application areas will also increase. It will be interesting to see how the market adapts to these changes.

What are the disadvantages?

Disadvantages concerning artificial intelligence tend to sound similar. It is often the idea that they are already so sophisticated that they are taking people’s jobs. Although partially true, it is not realistic in the foreseeable future. Instead of replacing humans in the workplace, artificial intelligence usage is supplementary. To use AI effectively, it makes sense for potential users to attend appropriate training courses which, in turn, involves time and costs.

Another disadvantage of AI usage in the MedTech sector is the lack of legislation, for the most part. There are only a few relevant regulations at a national or supranational level. However, these are gradually being implemented worldwide. The lack of laws is a disadvantage because their eventual implementation could significantly change the use of artificial intelligence. In such a case, manufacturers would have to adapt to the emerging regulations and possibly accept setbacks in development.

How is AI regulated?

The regulation of artificial intelligence concerning medical devices has been challenging to date, as there are still no standards and laws on AI. The decisive factors here are the fulfillment of requirements and the intended use of the manufacturer. It becomes obvious when observing the debate about whether ChatGPT is a medical device. According to the EU’s Medical Device Coordination Group (MDCG), software exceeds the threshold of a medical device because it has more than a mere storage and replay function. Therefore, a search engine is not considered as a medical device.

In contrast, ChatGPT could be categorized as a medical device. It can be used for diagnosis, monitoring, treatment, or disease prevention. So, it exceeds the functions of a pure search engine.

However, OpenAI has not registered its artificial intelligence as a medical device. It may also be due to the fact that it is not explicitly marketed, labeled, or seen as a medical device. To sum up, it is clear that the regulation of artificial intelligence is complex.

EU’s AI Act

The EU has recently worked on a cross-industry law to regulate artificial intelligence. The final draft of this AI Act was published in January 2024 and was confirmed on March 13th, 2024. It also explicitly addresses medical devices. Conversely, this means that medical devices and IVDs will have to meet the requirements of the MDR and IVDR along with those of the AI Regulation. Impacted areas of work include post-market surveillance, quality management systems, technical documentation, and risk management. It is unclear whether this might create a double burden for manufacturers and authorities or not.

Rule 11 in Chapter VIII of the MDR Annexes states:

Software intended to provide information which is used to take decisions with diagnosis or therapeutic purposes is classified as class IIa, except if such decisions have an impact that may cause:

  • death or an irreversible deterioration of a person’s state of health, in which case it is in class III; or
  • a serious deterioration of a person’s state of health or a surgical intervention, in which case it is classified as class IIb.

According to this rule, every artificial intelligence would belong to class IIa or higher and require a conformity assessment procedure. Irrespective of the actual application, they would fall under high-risk products in the AI Act. Requirements for them would increase, resulting in high costs for placing them on the market.

What needs to be considered?

The AI Act contains some duplication with the MDR and IVDR. It therefore remains to be seen whether medical devices with artificial intelligence will have to undergo double testing. The result would be a higher additional burden for manufacturers and authorities. However, as the artificial intelligence industry is developing rapidly, the AI Act may also be adapted and expanded in the future. Some discrepancies could thus be eliminated for the benefit of the authorities, manufacturers, and ultimately users.

[Disclaimer] This information is only one possible interpretation of the regulations. They are also in a constant state of change, so the information in this article may be incomplete or out of date. The above article is expressly no legal advice. Please refer to the official documents for information before making any business decisions. (Status of information: March 2024)

More MedTech News