I see this going 2 ways, people who see generative design as their overall design process, the "design by selection" approach to every element from beginning to end, diverging and converging, and those who see it as just part of the process as simply a divergence tool.
I wish I knew the answerer to this. Right now, the work processes are still evolving. Like CAD in the early days, no one imagined what it would end up like. Different designers are using it differently. I am looking at it more from the perspective of technology adoption. CAD drew very similar reaction from designers as it went from the labs in the 60's to general use in the 90's. These are the typical phases of any technology adoption.
1) Not noticed - laughed at and believed by a handful , no immediate practical applications.
2) Reaction - noticed, early adopters start engaging and non adopter oppose it vehemently.
3) Adoption - every one is using it and apparently nobody had anything against it.
In architecture generative design is in stage 3 with major practice using it and all major schools teaching it. In ID (judging from Core77 discussions) we were in 1 in 2005 and moving on to stage 2 now, given the current level or reaction, curiosity and engagement that we are seeing. So it is about 5 years behind Architecture.
Architects are not easily rattled by new ideas. For along time starting from the 70's (from the time of the architectural machine group at MIT ) there has been talk about partnering with machines in creative process and now, it is finally happening. In architecture I see different 3 different approaches.
1) Older and mature firms are using it for optimization (as in the bubble stadium) This fairly old stuff done by engineers routinely now used by architects.
2) Younger architects, using it for form and surface patterning exploration and using the outcome. This is mostly based on parametric design and the exploration is manually driven and is called generative design by most.
3) Truly generative design - where the form is generated purely by generative processes, this is still rare, but happening.
So we are seeing many different approaches. Interesting, some architectural educators see value in the challenging role of generative design as it challenges students to consider so many 'what if
' possibilities - purely for inspiration. I think these differences in approach and use will remain and that architects will be know for thier own styple of using generative design which will give their designs unique personality and character.