Regulating Tech – by Robin Hanson – Beragampengetahuan

Not sure exactly why, but one of its authors sent me the book Data Science in Context: Foundations, Challenges, Opportunities.

The chapter of most interest to me is ch. 7, on regulation, which highlights these recommendations:

4. Regulate uses not technology

5. Regulate clear, not potential, problems

6. Update laws with data science and technology in mind

7. Consider the impact of economic of scale and the virtuous cycle

8. Create independent, consensus-building institutions for difficult problems

Now points 4,5 are good solid advice that is ignored far too often. Such as re AI today. Yes, regulate the outcomes you care about, not the tech used to achieve them. And yes, trying to anticipate not-yet-realized problems just goes too far wrong too often. And points 6,7 are hard to argue with. 

But re point 8, I have more doubts. Sure when setting standards it makes sense to get the affected parties together to negotiate choices. But I worry much more when the main issue is how to protect customers and folks who can’t be well represented at the committee table.

Long ago as a junior employee in aerospace, I noticed that whenever some new issue or buzzword appeared in their world, ambitious folks would try to found or join a “working groups” on that topic. And such groups were eager to offer “recommendations”. Workgroup meetings didn’t take that much time, you didn’t need to be expert on the topic to join such groups, and membership gave you visibility and looked good on your resume. So such groups tended to offer recommendations that sounded socially desirable, while also benefiting their members. 

If you are going to regulate, you should probably listen to advice from such “consensus” groups. But maybe you should reject their recommendation to regulate in the first place.