Center for Strategic Assessment and forecasts

Autonomous non-profit organization

Home / Science and Society / Future of the Information Society / Articles
The future of health care: what prevents the introduction of artificial intelligence
Material posted: Publication date: 22-04-2019

Interest in the practical application of artificial intelligence (AI) is growing rapidly. Systems based on artificial intelligence, helping to improve the efficiency and increase the revenues of companies in various industries

However, despite the wide possibilities of neural networks, the spread of AI received not in all industries. Especially this trend is clearly visible in health care.

Today, artificial intelligence (AI) finds its application in many areas of our life — from transport and industry to Finance and medicine — and thus attracts more attention from investors. For example, in health care opportunities in AI now allow you to optimize the performance of hospitals, improve the effectiveness of aid delivery and make the diagnostic process more accurate. According to PWC, in the second quarter of 2018 artificial intelligence has become one of the most disruptive areas for investment in the United States with funding of $2.3 billion, Russia lags far behind the US and China for investment in the development of artificial intelligence, however, the popularity of AI in our country is growing. Despite this, the Russian doctors are ambivalent about the possibilities of AI in health care, pointing to the lack of infrastructure and uncertainty in the work. What barriers stand in the way of introduction of AI in the Russian health care system? And whether domestic doctors to find a common language with the new technologies?

To teach the AI

To artificial intelligence showed the most accurate results, it needs to be trained. This requires an array of well-prepared training data. Unfortunately, today in the health care of such information in the possession of experts in the field of Data Science a little, it is often incomplete, inaccurate, or "raw". The fact that the training data AI is provided to medical institutions before you provide the information they need to prepare (mark up to structure and digitize pathology, etc.). This process manual and time-consuming and costly. At the moment professionals involved in the preparation of these data in the health care system, is not enough. Is one of the challenges on the way to the creation of AI systems, able to produce results comparable with the results specialists. One of the main tasks in the field of AI today, which features programmers is the creation of Autonomous "intelligent" system self-motivated. While in medicine it is more widely used AI algorithms that were trained on the prepared data, algorithms, AI that can pass the training without the participation of a specialist, are still in the stages of development and testing. In these circumstances developers will need to work closely with doctors, the more the quality of the data, the more medical problems will be able to take on artificial intelligence.

To ensure data protection

Another obstacle for the introduction of AI in healthcare is the current status of the information security system. According to the report of the Infowatch company for the first half of 2017, medical facilities were leading in the number of incidents of data leakage (17,4%), and this trend is continuing. Earlier, in 2015, a striking example of the imperfection of this sphere was the theft of personal data of 80 million customers of health insurer Anthem. The incident not only caused damage to the users 115 million dollars, but also seriously undermined the credibility of the health system as a whole. Today, according to a study Philips "health future", in Russia 46% of citizens do not have confidence in those areas, where use of their personal information.

To solve this problem and strengthen citizens ' confidence in the security of their data, you need to look at it from two sides. First, we need to empower hospitals to exchange data on patients in the medical community, on the other hand to ensure better control and protection, so that patients could be assured that their personal information will not be available to the General public. This process is long and laborious, it requires the participation of both the state and business, but in the end, only he can really affect the security of sensitive patient data.

To comply with the rules of ethics

The second challenge concerns the ethical side. So far, we have not adopted the consensus — is it possible without threat to the health of the patient to use artificial intelligence to disease diagnosis and assignment of treatment and whether to refer the destiny of man "in the hands" of the machine. The fact is that with proper and high-quality training of the artificial intelligence is able to detect many pathologies, including malignancy and various pathologies at an early stage that will give more confidence in a favorable prognosis. But what if the patient has anatomy, or pathology is so rare that the AI with her never met? In this case, developers need to find appropriate new data, which will illustrate these features allow the machine to take them into account in new studies with a minimum error. And here from the experts there is a logical question: who is the true owner of the medical data patient, doctor, clinic, insurance company or someone else? And who has the right to dispose of them? But if AI makes a mistake and damages the patient? These moral issues are particularly concerned about the medical community and prevent the penetration of new technologies in health care.

What's next

Despite the challenges, the prospects for the development of artificial intelligence in medicine, there is. According to Deloitte, by 2020, world health expenditures reached $8.7 trillion to 10.5% of global GDP. Such forecasts put the government and business faced with the need to implement AI technology in medical institutions to provide doctors with new opportunities to improve the efficiency of use of existing resources and thus reduce costs. AI will cope with large data sets of patients and support physicians in clinical decisions. AI is especially popular in the field of radiology, pathology and genomics.

Thanks to the ability to collect and analyze the information that AI can increase the role of prevention of diseases of the population. For example, the Philips Research laboratory exploring the possibilities of artificial intelligence for early detection of tuberculosis. Experts have developed a solution based on AI, which analyzes a large number of medical images and makes conclusions about the presence of the disease. During research, scientists found that the results provided by the AI does not concede on quality to the work of a specialist.

To speak about full replacement of the doctor's artificial intelligence is premature. Today the developers claim — artificial intelligence should be treated as an effective assistant to the doctor, who adjusts to his work and helps make an accurate diagnosis the first time. It is likely that we will observe a symbiosis of computer and human, and the role of AI will rapidly increase as the enabling environment for its implementation. McKinsey experts claim that health is possible to automate 36% of the functions, particularly working with data. Therefore, we can safely predict that in the future the majority of routine tasks will be assigned to a machine, and doctors can focus on more important tasks that require human resource and professional approach.


RELATED MATERIALS: Science and Society
Возрастное ограничение