we don't know yet. it's not been legally tested.

It's not a question of if anyone holds the copyright, but a question of whether the models themselves and their outputs are derivative works of the data the models were trained on.

If they are derivatives, it may be possible that the licenses of that training data place additional restrictions on the generated code or forbid derviative works outright.

Speaking practically, for a random "FOSS" project that isn't very big, this is unlikely to ever come up. Courts would have to rule in favor of the training data owners, plus they would need to issue DMCA takedown requests and/or sue to have the content removed. That probably won't happen.

However, you should probably question any project where the developers don't understand the code well enough to write it themselves. I don't know if that's the case here, but that tends to be the case with vibecoded projects.

If nobody owns the copyright then it'd be public domain, which by definition is free software since you have the four freedoms with it.

The issue is that even that is legally dubious. Text generation programs have been shown to output chunks of copyrighted text verbatim, so it's possible someone actually does hold the copyright to parts of that code, you just don't know unless you do your due diligence on every bit of text that is output. Since neither the "author" nor the actual generator can "know" the provenance of the generated code it is very risky to assume it can be freely used.

if they didn't take the time to write it why should i take the time to read it?

it means it's legally dubious, it gives no guarantees about quality, and i don't want to use it.