AI chatbots use your conversations to train future models by default, and shutting that off is simpler than most people assume.
AI companies routinely feed user conversations back into their systems as training data, treating the prompts, follow-ups, and corrections people type as free labeled examples. That's standard practice across the industry, not a rogue exception. Opting out isn't buried behind some elaborate dark pattern - it's a legitimate setting, just one most users never go looking for. Finding it takes checking your account or privacy settings rather than assuming there's no choice at all.
The real story here isn't secrecy, it's inertia. Default-on data collection doesn't need to be deceptive to work - it just needs enough friction that most people never flip the switch, and that's precisely why so few do. That's a familiar pattern anywhere a company profits from your default behavior, from cookie banners to subscription auto-renewals.
So if training-on-your-data has been sitting in your assumptions rather than your settings menu, today's a fine day to actually check.