A client recently mentioned using AI between our sessions to research something we'd discussed — and I didn't mind at all. It got me looking properly at what the research says about AI and therapy: real evidence of benefit for purpose-built tools and between-session support, alongside genuine, well-documented risks around crisis handling, over-reliance, and young people in particular. Here's a balanced, evidence-based look at where it helps, where it doesn't, and why I'm not