(no subject)

[timelets]

When we talk about AI as a technology platform, the current discussion about alignment looks particularly deficient because it fails to address the enshittification problem.

Comments

Well ...

[ysabetwordsmith]

AI has started out shitty, so people may not think of it on the same trajectory as a product that starts out good but then begins to abuse people.

Re: Well ...

[timelets]

From a business perspective, enshittification is almost inevitable here because the system is going to know everything about users and will be in a position to shape their thinking. Unless some robust guardrails are installed, businesses will exploit this situation for profit and competitive edge.

Re: Well ...

[ysabetwordsmith]

The problem is that AI is being used as abusive, intrusive, and extractive from the start -- largely without protections. That makes it more of a threat than an asset.

[timelets]

I partially agree. When AI is deployed in an enterprise environment many employees can be treated as an expendable resource valuable only until the AI is trained and stripped of hallucinations. In terms of alignment, the tech is aligned with the bosses, not with the workers. It's a business asset.

We haven't reached this point in the consumer space yet. After all, AI providers need customers, gaining and keeping them forces some level of decency. But there are no guarantees whatsoever.