Table 4 Data extracts for sub-theme: Considering stakeholders’ requirements

Does this AI system really have your interest at heart? And, I think it's also sort of potentially concerning in all of those areas of healthcare where patients have quite legitimate views on which of different set of possible treatments they would want. Because I think it is a risk. So, that the AI system says that this is the best treatment (…). Yeah, but you could imagine AI advice being converted into slow pressure to take the best treatment, unfortunately. And sort of the fact that in many areas the best treatment is the one of the possible reasonable treatments that I want being sort of overlooked... [Rn5 (BE)]

I'm skeptical about rushing to use something that is automated just because we can. Just because we can, doesn't mean that we should(…). So, what's important for me is not that we have a product that works, but that product that reduces health inequalities that is able to work for the most disadvantaged people within the society, for those who typically struggle to access services in the first place... [Rn14 (DH)]

(…) Where we are sort of asking the doctors, like, uhm, "Okay, you try this out, like, from your perspective, like, what is the problem that you want to be solved?" Maybe we're not solving the problem that you want solved. We're solving the problem we wanna solve. So, like, what are the problems from your perspective? Like, how can we address some of those, you know. What, what would the system look like if it was useful for you. How would it fit in to your workflow? [Rn27 (BE)]

And in my eyes that has to really come from the clinicians that inputs, they have to come with problems and then people have to try to solve them and often, it's a bit the other way around it, that people with the solution, trying to find problems that they can solve. And that often leads to a lot of efforts and in the end, products that no one wants to use. [Rn29 (MEAI)]