@thomasfuchs I think a more and more common problem is that software teams are doing exactly that, but then assuming "we know what's best" for the people using that software. Or they end up using excessive amounts of AI to "solve" for the problem they introduced.