Click on the picture to purchase this issue

source

"Chinese Social Science Digest" 2024 Issue 1 P17-P18

Author's unit

Department of Science and Philosophy, University of Science and Technology, School of Public Affairs, University of Science and Technology of China, original title "Development trends in medical artificial intelligence ethics research - Theme distribution, knowledge basis and future prospects", excerpted from "Natural Dialectical Communication" Issue 12, 2023, Cui Jin excerpted

The ethical issues of medical artificial intelligence are gradually becoming a hot area of ​​concern to academics, industry and government health departments. Research on medical artificial intelligence ethics started late compared to technical ethics, but with the improvement of the maturity of medical artificial intelligence technology and the continuous expansion of commercial investment, the overall trend of medical intelligent transformation has become increasingly significant, and the attention to related ethical and moral issues has also increased rapidly. People are beginning to realize that the disruptiveness of medical artificial intelligence is not only an evolutionary iteration at the technical paradigm level, but also implicitly the possibility of innovation in the social system related to medical care. The development of research on the ethical dimensions of medical artificial intelligence is not only to deal with medical ethics review, but also to ensure that intelligent medical technology achieves positive social benefits.

Medical artificial intelligence ethics research focuses on the following contents in three different development stages.

1. The period of initial exploration. It can be mainly divided into three types of research directions: First, it emphasizes that ethics should become the constituent element of the theory of intelligent medical methods. Through the analysis of the current status of technology application, such as medical expert systems and electronic medical records, it is believed that technical ethical issues should attract the attention of relevant academic circles. This type of technical dimensions focuses on ethics, providing an endorsement of the technical necessity for the ethical research of medical artificial intelligence. The second is to explore possible response measures for medical artificial intelligence ethics. In the process of exploring technical characteristics and application cases, it is pointed out that the integration process of computing intelligence and life sciences requires attention to possible ethical issues. It is believed that artificial intelligence systems in medicine imply specific theoretical assumptions and social and cultural attitudes, and hopes to achieve the integrated development of medical humanities at the educational level through professional course construction. The third is to seek a knowledge management framework that supports ethical decision-making, and believes that the evaluation of medical artificial intelligence needs to be integrated with the clinical and social environment, as a decision-making auxiliary tool, providing participants with personalized and customized structured knowledge transmission, and providing doctors and guardians with necessary information support for moral decision-making.

2. Advanced surging period. The booming prospects of the medical artificial intelligence technology market at this stage have made ethical research pay more attention to practice-oriented ethical risk analysis: First, focus on technical ethical issues in different medical situations, and try to build an ethical and moral evaluation framework in the process of exploring the practice of nursing artificial intelligence; Second, deepen ethical cognition in the process of technology-based plasticization, and in the process of discussing application scenarios, related discussions or initiatives involving medical artificial intelligence and big data ethics.

3. A period of accelerated development. Research on the ethics of intelligent medical can be mainly divided into the following three main categories: First, the deep-level ethical issues of intelligent medical care, such as prejudice in injustice and algorithmic decision-making in health care, emotional out-of-control possible by human-computer interaction, trust relationships of deep learning in medical practice, moral dilemma caused by paternalistic behavior logic, and ethical challenges in the use of intelligent algorithms in psychological and mental health. The second is the construction of the ethical governance system of medical artificial intelligence, introducing theoretical frameworks such as stakeholder analysis and responsible innovation, emphasizing that the ethical practice of medical artificial intelligence must rely on reasonable evaluation guidelines, and technological innovation should proceed in parallel with ethical norms. The third is the value judging and epistemological basis of medical artificial intelligence ethics. It focuses on the reliability, interpretability and transparency of medical artificial intelligence computing process. It believes that doctors should have the necessary artificial intelligence skills and professional knowledge. Cooperating with medical informatics and data scientists is an important prerequisite for the ethical paradox of intelligent black box.

After nearly 30 years of development and accumulation, medical artificial intelligence ethics has gradually formed a special professional research field from the accompanying topics of technical research. The research form is no longer limited to the research on countermeasures of specific ethical issues. It has begun to explore the causes and backgrounds of ethical difficulties, and combined with the social system, it has accelerated its transformation to an integrated, systematic and inclusive medical artificial intelligence governance system. With the continuous acceleration of the iteration of medical artificial intelligence, the full-process medical artificial intelligence ethics system that runs through algorithm research and development, data acquisition, clinical practice, commercial implementation, health supervision and nursing is an important research topic to ensure the collaborative innovation of intelligent medical ethics and technology.

The knowledge base of medical artificial intelligence ethics can be divided into three categories: artificial intelligence technology, medical disciplines, medical big data and artificial intelligence ethics.

The practical domains of the development and application of each subdomain of medical artificial intelligence are very different. Therefore, medical artificial intelligence ethics should be divided into two major modules: general ethics and subdivided ethics. The focus of the discussion on general ethical issues should be universal ethical difficulties, such as the risk controllability of data prediction, data platform permission configuration, algorithm bias in machine learning, and fairness in the allocation of intelligent medical resources. Related research on artificial intelligence ethics basically started here. The next step is to clarify the inducing mechanism of general ethics. Is it a lack of conditions at the technical principle level, or is it an exogenously driven alienation of technology in the social environment? What is the weight of the endogenous technology and external environment on the shaping of the ethical form of medical artificial intelligence? All need further analysis and judgment.

Subdivided ethics requires specific analysis of ethical issues that may arise in different medical reality contexts. For example, in the face of the complex mechanisms of human health, can the health prediction provided by smart wearable devices ensure accuracy? How to overcome the possible emotional panic and excessive anxiety? Introducing intelligent technology into patient diagnosis and treatment and hospital management, the dominant position of doctors is at risk of being eliminated by algorithms. How to ensure the humanistic care of traditional medical ethics? Will weakening the parent-style diagnosis and treatment model and increasing the autonomy of patients and consumers cause more difficult responsibilities? The moral dilemma that this series of medical artificial intelligence segmentation technologies may arise are key issues that need urgent attention and research.

The most important special feature of the ethical issues caused by the deep intervention of artificial intelligence in the field of medicine lies in the substitution of intelligent artificial objects on the dominant position of doctors. At present, medical artificial intelligence ethics has begun to enter a stage of accelerated development. A professional research system is being promoted simultaneously, and a series of medical artificial intelligence ethics standards have been proposed, a large part of which are derived from traditional medical ethics principles. However, due to the transformation of medical subjects, the changes in interaction patterns, and the transformation of medical information forms, it is necessary to re-examine the applicability, reliability and completeness of traditional medical ethics norms. The "de-humanization" of the transformation of medical intelligence has made the risk of "breaking" with traditional ethical concepts in the context of the accelerated commercialization process. Therefore, the next stage of research on the ethics of medical artificial intelligence needs to deepen innovative interpretations of medical artificial intelligence ethics and social issues based on medical ethics and technical ethics theory. On the one hand, it is necessary to subdivide the ethical application standards for the implementation of technology based on different medical scenarios. On the other hand, it should also target holistic ethical dilems, such as the establishment of the quasi-moral subject status of intelligent medical care, and how the medical humanistic spirit can survive in the era of intelligence, and make ethical response preparations. The development of these research is not only a reflection of the continuous advancement of artificial intelligence ethics research, but also can feed back to the theoretical system of traditional ethics and accelerate the innovative development of ethics itself.

From the perspective of the social shape of technology, ethics and morality are the necessary non-human actors to build a medical artificial intelligence society-technical system. It should be made clear that medical artificial intelligence ethics and ethics are necessary supporting modules for intelligent medical technology and require joint consultation and participation of stakeholders such as ethicists, artificial intelligence scholars, technology enterprises, and government agencies. The ethics of medical artificial intelligence have gradually shifted from simple academic issues to practical research on policy norms and management systems in intelligent medical practice. Medical artificial intelligence ethics is not only the establishment of technical norms and the education of researchers' literacy, but also an important theoretical basis for medical artificial intelligence laws, management and policies. To achieve the corresponding matching of medical artificial intelligence ethics and governance system, not only should we make up for the ethical review mechanism of medical artificial intelligence clinical and product, but there are also many transitional ethical issues that need to be solved urgently. For example, in the strategic layout of artificial intelligence, what is the strategic positioning of medical artificial intelligence ethics? How to embed ethical value in medical artificial intelligence technology design through legal and industry guidelines? How can the ethical and institutional levels be coordinated with medical artificial intelligence innovation? Implementing the two-way support of the ethical norms of medical artificial intelligence and the social governance system is an important external synergy factor to ensure the reliability and social feasibility of medical artificial intelligence technology, and it is also an institutional guarantee for preventing the derailment and alienation of medical artificial intelligence technology goals and practices. This series of intersectional issues still requires further attention and research from the academic community.

In order to ensure the review and normative effectiveness of medical artificial intelligence ethics, closed social experiments are needed as intermediary links. Through a series of pilot feedback, the guiding principles and regulations of medical artificial intelligence ethics are continuously adjusted and optimized. Practical achievements that define the effectiveness of medical artificial intelligence ethical standards come from social consequence evaluations within limited boundaries on the one hand, and on the other hand, we need to pay more attention to the psychological and emotional feedback of the technical audience. Psychological research can provide an empirical reference to the patient or healthy consumer perspectives for the practical evaluation of medical artificial intelligence ethics, and provide practical evidence support for technical ethics. The construction and improvement of the ethical norms and governance system of medical artificial intelligence not only requires the "top-down" ethical principle as a framework regulation, but also needs to consider the "bottom-up" ethical demands of various stakeholder groups. The ethics of medical artificial intelligence cannot be simply regarded as a mandatory constraint. It also needs to pay more attention to the emotional feelings of the technical audience and test the practical effectiveness of ethical norms. Psychological methods can analyze and judge the possible direction of technical adjustments through changes in individual psychological acceptable levels, and can provide psychological evaluation based on ethical practices of specific medical scenarios for medical artificial intelligence ethics, and should become an important auxiliary to the next stage of medical artificial intelligence ethics.