Table 1 Summary of the stakeholders’ interests by category
Stakeholders’ common and specific points of view about AI | |||||
|---|---|---|---|---|---|
Stakeholders | Health professionals/Physicians | Health industry | Individuals without conflict of interest | Regulatory agencies | Health researchers |
General | |||||
Common notions for all stakeholders | Fuzzy notion of AI—Issues around health data—International competition—Development of AI—Change in healthcare relationship—Radiology as a precursor | ||||
Specific | |||||
Priorities | Give the best care to patients | Be efficient | Protect individuals and their rights, but also improve health | Regulate appropriately (legislating or providing guidelines) | Generate research results |
Driving forces | Will to integrate these tools into practice | General will to develop AI | None | Omnipresence of the subjects (some actions have already been led) | Presence of encouraging results (due to big data and machine learning) |
Possibility of a change in medical training | State of the art of AI in healthcare | Development of AI’s applications | Existence of active work on AI | ||
Points of vigilance | Waiting for proof in current practice | None | Promote population’s education/information | Not to succumb to the ambient willingness to legislate | Not to call everything “AI” |
Promote population’s informed opinion | |||||
Respect privacy | |||||
Evaluate issues of social justice | |||||
Obstacles (scientific and/or legal) | Questions of liability/responsibility | Questions of liability/responsibility | Questions of liability/responsibility | Not feeling able to evaluate AI’s software yet | Financial context: need of funding |
Misreading of institutional support | |||||
Regulatory context | Difficulty to entirely understand the subject | ||||
Difficulty to access to annotated health data (need for professionals) | |||||