When you use AI to discuss your medical history across multiple conversations, it starts fresh each time unless you paste in context—meaning it might give you contradictory advice or miss important details from earlier. Knowing this limit helps you decide when to save a complete summary of your situation for the AI, rather than expecting it to remember what you mentioned yesterday.
You're explaining your medical history to an AI—you start with childhood eczema, mention your teenage asthma, describe your current joint pain, talk about your medications, list your family history. Halfway through, the AI suddenly seems to forget you mentioned you're allergic to penicillin. That's not carelessness. It's a technical limitation called the context window.
A context window is the amount of information an AI can actively 'hold in mind' at one time. Think of it like your working memory: you can juggle five balls at once, but if someone hands you ten, you'll drop some. Each AI model has a maximum context window (measured in 'tokens,' which roughly equal words). Once you exceed it, the AI starts losing earlier information.
Most modern AI tools have context windows between 4,000 and 100,000 tokens. That sounds like a lot—roughly 2,000-50,000 words—but medical conversations can eat through that quickly when you include background information, medication lists, test results, and detailed symptom descriptions.
Your medical history is interconnected. Your current symptom might be explained by something you mentioned twenty exchanges ago. If the AI's context window has shifted, it might make suggestions that contradict information you already provided. This isn't AI being dumb—it's AI hitting a real technical boundary.
You'll notice the context window is affecting the conversation if AI starts repeating questions you already answered, suggests treatments that contradict information you provided, or seems confused about your medical history.
Try this: Have a medical conversation with an AI tool and deliberately make it long—ask 40+ questions across 15+ exchanges. Toward the end, ask the AI to recall something specific you mentioned at the very beginning of the conversation. See if it remembers. If not, that's your context window at work. Then start a new conversation and paste a one-sentence summary of that early detail. Notice how it performs differently.
Peri can explain this concept, give practical examples, help you decide whether it applies to your situation, or recommend a journey if appropriate.
Explore related journeys or tell Peri what you're working through.