When there's intelligence adding a will should be trivial. You just tell it to do something and give it some actuator, like a web browser. Then let it run.
Not that I agree with the trivial part, but that’s a good question. Afaiu, current AI has a “context” of few thousand something, in which it operates. If someone enlarges it enough, loops it to data sources, makes its output to do real things (posts, physical movements), then it’s a stage for “will”. You only have to prompt this chain with “chase <a goal> and avoid <destruction conditions>”. If we humans didn’t have to constantly please thermodynamics and internal drives, we’d stay passive too.
