Oh with regard to this one, I don't think my wording was clear. I meant with regard to testing or challenging to be certified without having a formal CS/related background. As: a doctor would have to have an MD to practice as a doctor amongst other certifications— I was contending that an explicit CS degree may not be an optimal equal designation for practicing software engineering/research, as it were. As a background— there are many talented and influential researchers who would be cut out of practicing if the line were drawn at a reputed CS degree. Aligning "software practice" (for lack of better wording) hard with a CS degree might be poor bounds for the field.

But I definitely think I agree, at least on a high level with what you're proposing. I hadn't considered it. Good things have come out of research using the large amounts of data available, so it should continue. But there definitely should be some sort of bounds and method for accountability. Also would include a special permissions and appeal process. There's I'm sure a lot of cost/benefit judgement as there is in many scientific experiments (there's seemingly a great deal of that in biological testing).

And you also might restrict certain entities from performing the research and instead be compensated for their collected data by a reputable research group. Said group can produce the hard/applied research, patents, and license them to groups to use them.

Maybe this would also undo some of the effects of outsourcing/offshoring US coding practices due to the need for ethical compliance (at the very least in mission critical systems e.g. vehicle software, hospital software, etc)