인공지능 기반 진단방법 특허에 관한 연구

Abstract

Humanity has been striving for health and life extension for a long time. and early diagnosis of the disease is of paramount importance. However, there are labor, geographical and environmental limits to receive sufficient medical services. In particular, it can be said that there are cases where medical demand explosively increases in a short period of time due to Pandemic Diseases, and that it is an important problem that our society faces in that it is gradually entering an aged society. In the midst of such difficulties, AI technology, which is rapidly developing recently, is attracting much attention as a means of accurately diagnosing a large number of patients in a short time. However, the AI-based diagnosis method under the Patent Act has problems because it contains issues related to medical methods and software, which have been difficult discussions for a long time. But in terms of medical methods, our patent examination guidelines have recently been revised to recognize industrial applicability only for diagnosis using digital means among medical methods. In terms of software, patent examination guidelines in each country are being reorganized to promote technology development in the field of AI. However, contrary to expectations, there are not a few cases that are rejected in relation to the above two issues. Therefore, in this paper, I would like to review and present the matters to be noted in order to obtain a patent for an AI-based diagnosis method. First of all, in terms of the diagnosis method, it should be clarified that the subject of Clinical Decision Making in all stages constituting claims is digital means such as AI, not human beings. In addition, even in some steps, ambiguous expressions that may involve humans, such as Self-Diagnosis, should be excluded from the claims. Furthermore, it is desirable to describe the claims in the active form of AI, and due to the time-series nature of method invention, attention should be paid to the step-by-step connection relationship. Care must be taken not to describe steps that suggest or imply prescriptions or treatments using the results after the AI analysis step. Next, we need to be careful in terms of AI. Recently, for the purpose of industrial policy to promote technology development in the field of artificial intelligence, revisions of patent examination standards have been carried out in each country, but in principle, Machine or Transformation are required to be patent-eligible for an AI-based diagnosis method. Accordingly, it would be desirable to use a device for information exchange with humans, an accelerator which is generally developed together with AI, as a component of the claims. Finally, it should be taken into account that it may be against public health. This is because even if the false diagnosis rate of AI is quantitatively lower than that of human doctors, qualitatively more dangerous cases may result if analysis errors are caused at a basic level. However, there is an aspect that it is difficult to see that AI-based technology directly harms public health according to medical law article. And since benefits and disadvantages coexist in all technologies, it is necessary to weigh the benefits. Furthermore, it should be fully considered that AI can function as an effective means in the recent difficult health environment.