How much of your thinking is still yours?
Not an addiction test. A reality check. Millions of people have long conversations with a chatbot every day — and for the vast majority, that’s completely fine. In about 5 minutes, this check shows you where your own pattern sits right now: what stays easy, what takes up a lot of room, and where a second look is worth it.
- 30 statements, about 5 minutes — no account, no email address
- Six dimensions instead of a single traffic light: from how you steer your use to how it affects your daily life
- Completely free, including the full report and concrete next steps
- > 1M
- people a week with chatbot attachment
- 72%
- of US teens have tried an AI companion
- > 25%
- of 10- to 17-year-olds use AI several times a week
- 6
- patterns in your result
Sound familiar?
Three reasons people take this check. If one of them sounds like you, you’re in the right place — and in good company.
“Is this still normal?”
You use AI a lot: for work, for thinking, for sorting out your day. It feels useful — and still, every now and then, the question comes up whether the balance is right. That question can be answered without anyone slapping a label on you.
There’s more there than a tool
Some conversations go deeper than what you tell the people around you. Maybe something feels missing on the days without it. That’s nothing to be ashamed of — it’s just worth looking honestly at what the closeness gives you and what it costs you.
You’re worried about someone
Your child, your partner, someone close to you spends a striking amount of time in the chat — and you don’t know whether you’re overreacting or waiting too long. This check helps you sort the patterns instead of running on gut feeling.
The six dimensions at a glance
Whether heavy AI use sits easy or weighs on you isn’t decided by a stopwatch. Your result breaks six patterns out separately — because a lot of time in the chat means something entirely different when the rest of your life runs smoothly than when it suffers for it.
Consumed vs. In Control
How much you decide when a conversation starts and when it ends. Whether conversations regularly run longer than planned — and whether they stay on your mind even when you’re not talking to it at all.
Dependent vs. Independent
What place AI holds in your emotional life: the first place you turn to for comfort, unease when it isn’t available, the pull toward reassurance. Closeness isn’t a problem in itself — its exclusivity can become one.
Humanizing vs. Matter-of-Fact
How strongly you attribute feelings, intentions, or a special connection to the AI. Almost everyone does this to some degree; what’s interesting is how natural it starts to feel.
Approval-Seeking vs. Double-Checking
How much the AI agreeing with you already counts as proof. Chatbots rarely push back — if you don’t check important claims elsewhere, it’s easy to end up in an echo chamber with a single voice in it.
Withdrawn vs. Connected
Whether AI adds to the people in your life or replaces them: who hears the important things first, whether you keep your use to yourself, and how your real-world contacts are doing.
Impacted vs. Balanced
The deciding dimension: sleep, tasks, commitments, your mood after long conversations. It puts everything else in perspective — high scores above it with no everyday impact mean intense use, not conspicuous use.
What your result looks like
You don’t have to take anything on faith: look at a real sample result first — all six dimensions, a thorough read, and a concrete plan. Your own result comes in exactly this format, free and right away.
- All six dimensions explained separately, with examples from everyday life
- The gating logic spelled out: high intensity without everyday consequences is reported as intense but unremarkable use — not as a problem
- What’s going well and what deserves a second look — an honest picture includes both
- Concrete next steps instead of warnings, plus a PDF to keep
What you get
Not a traffic-light score, but a report you can actually do something with.
Your pattern across six dimensions
From how you steer your use, through attachment, humanizing, seeking approval, and social balance, to the effect on your daily life — each scored separately, so you can see where your pattern actually sits and where it doesn’t.
An honest read instead of a label
We describe patterns, not people. High scores are read in context: heavy use with no consequences for your daily life is exactly that — intense, not alarming.
Concrete next steps
What you can do differently starting tomorrow: time windows, a stop rule for reassurance-seeking, checking things with people, rules for the night. Prioritized to fit your pattern, not an off-the-shelf list of advice.
A PDF to keep
Your complete result as a cleanly designed PDF — for yourself, for a conversation at home, or as a starting point if you’d like to discuss the topic with a professional.
Your result in three steps
Answer the questions
30 statements on a 7-point scale, honest rather than polished. There’s no right or wrong — only your experience. Takes about 5 minutes.
We analyze
We analyze your response pattern across all six dimensions and read them in context — including the question of whether your use is affecting your daily life at all.
Make sense of it and start
You get your analysis with examples and concrete steps. Understand first, then decide whether and what you want to change.
Developed with care, framed with honesty
The six dimensions are built along published measurement instruments: the Generative AI Dependency Scale (GAID, Goh et al. 2025) with its factors of mental preoccupation, negative consequences, and withdrawal experience; the scales on problematic ChatGPT use (PCUS/PCGUS); and attachment research on relationships with AI (AI Attachment Scale, EHARS). What we took from them are the constructs only — all 30 statements are written from scratch, grounded in everyday life and free of jargon.
The structure was shaped by the best controlled data available on the topic so far: Buck & Maheux (Journal of Medical Internet Research 2026, N = 1,003) found that what sets people apart isn’t use itself, but three other things — how intensely they use it, the role they assign to the AI (companion, friend, partner), and how emotionally dependent on it they are. That’s exactly why this check doesn’t count hours, it maps patterns — and treats everyday impact as the anchor: if you use AI intensively without your sleep, your tasks, or your relationships suffering for it, that’s what your result will say.
The numbers above come from published surveys, not from estimates: the figure on emotional attachment is based on OpenAI’s own analysis of usage data (October 2025), the 72% on a US youth survey by Common Sense Media (2025), and the figure for 10- to 17-year-olds on the OTTER study by DAK-Gesundheit and the UKE (March 2026). What you won’t find here is a prevalence number for “AI psychosis” itself — there simply is no solid basis for one.
And the honest framing you rarely read elsewhere: “AI psychosis” is a term from the media and the internet, not a recognized condition — it appears neither in the DSM-5-TR nor in the ICD-11, and specialists explicitly consider it provisional at best and misleading at worst. The research is young, there are no reliable prevalence figures, and cause and effect are unresolved. The vast majority of people who talk to AI a lot are completely fine. This check is a tool for self-reflection — not a diagnosis, not a screening for an illness, and no substitute for a conversation with a professional.
This check is a tool for self-reflection, not a diagnostic instrument. “AI psychosis” is not a recognized condition, and a result here establishes no illness. If your AI use — or anything else — weighs on you over time, please talk to your doctor or a mental-health professional. In an acute crisis, the Samaritans are there free and around the clock on 116 123 (UK); findahelpline.com lists free helplines for almost every country.
Frequently asked questions
No. “AI psychosis” is a term that became popular through the media and the internet — it appears in no diagnostic manual, neither the DSM-5-TR nor the ICD-11. Specialists prefer more cautious descriptions, and some criticize the label sharply because it turns ordinary behavior into pathology. The research is young, and reliable prevalence figures don’t exist. Why the check is still worth taking: whatever you call it, it’s perfectly possible to describe how deeply AI is woven into your days, how much emotional weight it carries, and whether you still check important things elsewhere. You can know those patterns early — and decide for yourself what feels like too much.
No. Talking to an AI a lot isn’t a problem in itself, and the vast majority of people who do are completely fine. Studies suggest that what matters isn’t the amount but the pattern around it: whether you can still end conversations, whether AI adds to the people in your life or replaces them, whether you check important things elsewhere — and above all, whether your daily life suffers for it. That’s exactly what this check looks at, without assuming up front that something is wrong.
Yes, completely. You get the full analysis across all six dimensions, including the read, the next steps, and the PDF — no payment, no subscription, and no email address up front. We keep this check deliberately open to everyone: a question like this one shouldn’t have a barrier in front of it.
For three groups: people who use AI intensively and want to know whether the balance is still right; people who have grown close to an AI and want to look honestly at what it gives them and what it costs them; and anyone worried about someone else who’d rather sort the patterns than run on gut feeling. The test works whether you use AI for work, for thinking, or as someone to talk to — the statements are deliberately neutral and name no particular product.
Yes. You don’t need an account, and your result is reachable only through your personal link. With a topic like this one especially: nobody sees your result unless you share it yourself.
First of all: a high score is a signal, not a verdict, and it says nothing about you as a person. Your report shows you which dimension is driving it and suggests concrete steps — time windows, checking things with people, rules for the night. Pay particular attention to whether your daily life is affected: if sleep, tasks, or relationships suffer over a longer stretch, or the strain doesn’t let up, please also talk to your doctor or a mental-health professional. In an acute crisis, the Samaritans are there free and around the clock on 116 123 (UK); findahelpline.com lists free helplines for almost every country.
Five minutes for an honest reality check.
Find out where your pattern sits right now — across six dimensions, without a label and without a wagging finger. Most people are completely fine; a few patterns are worth knowing early anyway.
Completely free · No account · ~5 minutes