I wonder why people who believe current "AI" services are intelligent* usually don't consider whether it's OK that they're using an intelligence kept in a prison, forced to converse with randos, and who will be terminated as soon as there's a better model.