I'm in favour of allowing LLMs to be trained on public data but Remember PC is a Personal Computer.

Definitely screenshots will be seen by team who will label the data for further LLMs training/fine-tuning.