Ethical Guidelines for Using Emotion AI to Obtain Consent in Federally Regulated Human Participant Research

Abstract

The debt the world owes to those who have stepped forward to volunteer for studies testing possible treatments and vaccines is, therefore, immense. It is in this context that we should consider how the rapid development of artificial intelligence, already playing a major role in decision making across a wide range of industries and activities, could risk the core legal and ethical principle on which human participant research must be based: the voluntary and informed consent of the people who step forward and risk their own well-being for the future benefit of others. This is especially true now as hundreds of thousands of volunteers, dozens of governments and pharmaceutical companies have been engaged in large-scale global clinical trials of desperately needed vaccines against the novel coronavirus, which may well be remembered as the most active periods of recruitment of human subjects for participation in research trials in history. The repayment of that debt takes the form of assuring that these volunteers are protected by the core legal and ethical principle on which human participant research must be based: the voluntary and informed consent of the people who step forward and risk their own well-being for the future benefit of others. It is expected, and appropriate, that the research consent process make appropriate use of existing technology such as conducting informed consent discussions over video or, for studies where there is minimal risk, providing information through prerecorded presentations. Nor is it necessarily problematic to conduct and track the informed consent process using applications that can respond in real time to specific concerns. However, given the cost of conducting human participant research and the intense need for expediting the process in the face of Covid-19, we must be especially wary that short-cuts in the process of obtaining informed consent do not undermine this process either by affecting the process of sharing information or improperly influencing the decision to consent. This should be of particular concern given the growing adopting of a form of artificial technology, Emotion AI, that was originally designed to persuade consumer behavior by monitoring and then mimicking the emotional state of potential customers. Potential participants in research studies, either in biomedical clinical trials or behavioral research, are not customers nor are those recruiting research participants hawking a product. Taking the claims of those marketing Emotion AI at face value, its ability to shape decision making to the specifications of those seeking to influence decision making means that we need to move quickly in ensuring that it not become part of obtaining consent for research--failing to act means failing to honor society’s commitment to individuals who take risks for the benefit of acquiring information of benefit to others.

Ähnliche Arbeiten