Of course there is value in promulgating safety *guidelines*.

But we cannot guarantee those guidelines to always be followed.

Sure, and we can’t guarantee you’ll read the safety instructions that came with your chainsaw. That’s orthogonal to the questions of whether those instructions should exist, whether “power tool safety” concepts should ever be promoted in society, and who’s ultimately responsible for the use of a tool.

Absolving humans of all responsibility for the negative consequences of their own AI misuse seems to the strike the wrong balance for a healthy culture.

Guidelines on their own probably won't be taken too seriously.

But other things will:

- Liability rules

- Regulations that you get audited on (esp. for companies already heavily regulated, like banks, credit agencies, defense contractors, etc)

If you get the legal responsibility part right, then the education part flows from that naturally.

Notwithstanding that the guidelines will even be applicable in the quiet versions that get deployed when you aren't looking. It's a constant moving target, and none of the fanboys will even acknowledge the lack of discipline in it all. It's fucking mad. And I say this as one who can see utility in the tools. But not when they are constantly shifting their functionality and behaviour.

One day everything works brilliantly, the models are conservative with changes and actions and somehow nail exactly what you were thinking. The next day it rewrites your entire API, deploys the changes and erases your database.

If only there was intellectual honesty in it all, but money talks.