Users on SO created value and freely shared it with a community in expectation that the value they created would be freely and collectively shared with everyone. In SO's case this expectation was explicit; the data backup and API was billed as a deliberate choice designed to give users the freedom to migrate and scrape data in case the company went "evil." It was designed specifically to reduce SO's ownership claim over user-generated content.
It's not that SO has a moral right to control and profit from that content. The reality is that SO holding that content at all is a conditionally granted privilege that the community affords the site, and it is a privilege that was always designed to be revocable and the data moveable if SO started abusing its position of power as a host and trying to lock down access.
Some writing/content sites that have taken steps to restrict AI access based specifically on community request. That's a very different situation; if a community (particularly a closed or close-knit community) is collectively and (mostly) uniformly trying to avoid an AI scraping the content that they created, then good for them. There are communities online that are in that position. But "how will the company get reimbursed for our valuable asset" should not be part of that conversation. And SO in particular was set up around norms that deliberately allowed this kind of scraping. It's not their asset to protect.
> rent-seeking AI models
I have issues with modern AI economic models too but I don't think that "rent-seeking" is an accurate term to use. A better word would probably be "parasitic"; I understand (and at somewhat agree with) the argument that OpenAI is looking to repackage information it didn't create in a way that redirects attention away from the original source of information.
But I'm having a really hard time figuring out how OpenAI is hoarding a scarce asset to extract value by controlling access to that asset. The more obvious rent seeking behavior here is coming from SO, a company trying to restrict access to Creative Commons licensed content created for free by unpaid volunteers, and trying to reclassify that content as their corporate property.
I guess being as charitable as possible, I do worry about the SaaS model of many AIs that are dedicated to content generation, and I worry a little bit about AI models becoming heavily integrated into creative processes and then extracting a kind of monetary "creative tax" from artists/creators while heavily restricting what they are allowed to make. That's at least adjacent to rent seeking, but I'm still not sure it's the term I would use and I'm not convinced it's a scenario that's applicable here.