Humans see and hear, we make plans, and we adapt to change. Artificial intelligence (AI) is when we replicate such cognitive performances on machines or computers. However, we still don’t know in detail how our brains do these things. But we can think of mathematical procedures that perform similar tasks in certain areas. With algorithms, machines can use sample data and derive models from them to improve their behavior step by step. From chess or other brainteasers on the Internet, the duel of man against machine has been one of the classics for years. On the one hand, it fascinates us to see how computers and machines learn on their own and adapt their behavior without human intervention. They are getting better and better and are already superior to us in some cases – just a few days ago, a computer beat five poker pros at once for the first time. This potential must be exploited, also in the field of medicine, where many treatment methods are becoming better and better through the use of intelligent software. Whether it’s apps for the early detection of diseases or personalized cancer therapies: Intelligent systems are expanding the possibilities of the medical profession quite considerably. On the other hand, there are still some challenges to be overcome, which AI brings with it. Is it ethical to listen to a machine in sensitive matters of life and death? What framework do we provide so that technology always serves people – and not the other way around?
What is AI capable of?
Self-learning algorithmic systems do nothing other than independently search for patterns in a huge pile of data that humans would not recognize or would recognize only with the greatest effort. Therefore, such learning systems are particularly suitable for repetitive tasks such as searching for anomalies, deviations, or commonalities and can also generate meaningful results from new data, i.e. they do not have to be reprogrammed every time. For example, in computed tomography scans, which doctors have hundreds of in hospitals every day, or blood cancer diagnostics by evaluating blood samples using flow cytometry after hours of gating. Through appropriate software development, AI technologies provide valuable services in medical diagnosis and therapy.
Where are the opportunities for AI in medicine?
AI brings enormous advantages. First, they can recognize relevant patterns that humans might never have looked for in the first place. Moreover, unlike humans, they never get tired or frustrated. They work without rest and can theoretically be fed with more data indefinitely because they learn at will. AI algorithms thus relieve humans of the very activities that dehumanize them. Numerous studies impressively prove that AI systems are very good at this and – for example – can reliably evaluate images. A study from Stanford, for example, showed that self-learning algorithms can classify skin cancer as competently or even better than dermatologists. Another field of application is personalized medicine. Here, for example, gene expression data is evaluated on an AI basis to arrive at tailored therapy recommendations. AI is also making inroads in the field of robotics. Medical robots or nursing robots can benefit from AI and become more autonomous. To improve healthcare, increasing patients’ chances of recovery, and supporting doctors in their diagnoses and therapy decisions, AI will increasingly become a part of healthcare in the future. The basic prerequisite for AI applications is data – including patient data, for example from the planned electronic patient record.
What are the prerequisites and framework conditions for AI?
The technical and organizational conditions required for the quality-assured use of AI assistance systems in medicine include the certification of AI systems and access control mechanisms to protect against attacks, as well as the integrity of data records and secure transmission paths. There are still several potential hurdles to overcome on the road to the widespread use of AI medical devices. First and foremost is a secure and powerful IT infrastructure for the storage and transmission of healthcare data and the digitization of care processes. To make this data accessible, it is necessary to establish suitable care registers for research and development purposes. There is a considerable need to catch up in this area in Germany. It is to be hoped that the laws currently on the way, above all the Hospital Future Act, the Digital Care Act, and the Patient Data Protection Act, will have the intended effect.
What are the prospects for AI?
Despite the above-mentioned challenges, it is already becoming apparent that artificial intelligence has become a key technology. It will help to overcome the current challenges facing the healthcare sector. Above all, AI ensures that diseases are detected more precisely and earlier. The quality and affordability of medical care should benefit from this.