A sensible choice. Now only if open source developers would update their licenses, perhaps a new GPL license, to restrict reselling of IP through AI models. These folks need to adhere to rules if we are to have a healthy ecosystem.
The right thing for them to do morally, would be to implement content visibility/privacy controls for their users similar to what Facebook offers (strange feeling to be referring to Facebook in this context).
Basically what I want is that all models trained on open source data or user created content without proper licensing are also open source and free.
(Also when the Twitter APIpocalypse happened, which the article forgot to mention.)
If that holds up legally, the best you can do is to try to stop your content from being scraped or not release it at all.