OpenAI forms safety committee as it starts training latest artificial intelligence model

Any organization which establishes its vision/mission/culture and then, much later forms a committee about safety..

isn't going to change.

the committee is a "bandaid": cosmetic, only.

Organizational-culture is set right at the beginning, & to change it, you have to do a grass-roots campaign, that includes everybody, until the grass-roots motivations themselves force-turn-the-tide.

This is 95%-certain to be appearances-only.

iirc, GE did an actual culture-change accomplishment: their CEO set up a system where the low-level managers went on retreats/workshops, and the result of those workshop/retreats was that they had to come up with a way of improving things in their part of the company..

( the problem this program was fighting was the top-down culture-of-authority-and-no-adapting )

Their boss had to OK or disallow their solution on the spot, when they got out, I think..

Once enough percentage of the low-level managers had been through the process, then the tide turned, & became a tsunami of change, wiping-out all obstacling before it.

Notice the difference, though:

a committee won't produce culture-change.

Grass-roots can, but it has to be done correctly.

_ /\ _