Q&A: Responsibility for bias, lawsuits and privacy concerns

During the webinar Q&A, attendees raised practical policy and liability questions. Tom Petros asked whether responsibility for bias belongs with model trainers or with human prompt writers; Gadeau replied it is both "things you can control and things you can't," urging individual users to practice responsible prompting while noting that companies must improve model design.

On legal accountability, Gadeau said "there's an awful lot of lawsuits against AI companies right now" but did not identify a single high‑profile bias suit off the cuff. On privacy, he confirmed that feedback features (thumbs up/down) "probably" influence future training and that free models often contribute user queries to datasets, advising caution about entering PII into prompts.