The adoption of #ai in coding is very concerning, especially in free software projects. If it's okay to train on free software, it's also okay to train on proprietary software. Otherwise, it's a nonsensical double standard: in both cases, the licenses have terms that must be adhered to. That projects and companies ignore this out of convenience is very worrying.
There's also the negative socioeconomic affects on people + planet: big data centers consuming massive amounts of energy and water.