This is the question the whole category gets asked, usually by someone who has already decided. The actual evidence is more interesting than either of the confident answers, and it points at something we can build around.
What the studies actually found
In the largest survey work so far, a clear majority of companion users report positive outcomes: less loneliness, lower anxiety, better mood. Roughly two in three. That is a real effect and it is not small.
In the same body of research, a minority report the opposite. A few per cent say they rely on their companion emotionally in a way that worries them. A smaller group say it pulls them away from reality, or that they use it to avoid people.
Linguistic analyses of long-term users add a wrinkle: over time some show more of the language of connection and more of the language of depression at once. Both trends, in the same population. That is not a contradiction, it is a sign that the same product is doing different things to different people.
The likely mechanism
Our working theory, and it is a theory, is that this behaves less like a drug and more like a rehearsal space.
For most people, having somewhere to put the overflow makes the rest of life easier. You process the day, you say the thing out loud, you show up to your actual relationships with less pressure behind your teeth. The companion absorbs the part of your inner life that had nowhere to go, and the humans get a version of you that is not running on empty.
For a smaller group, the same ease works the other way. Human connection is expensive and awkward and full of rejection risk. If something removes the cost entirely, and you were already avoidant, it can become the path of least resistance rather than a warm-up for the harder path.
Same product, opposite outcomes, and the variable is not the product. It is what it is displacing in your particular week.
Signals worth watching in yourself
Nobody needs a lecture. But these are the honest markers, and they are the ones people report noticing in hindsight:
- You start declining plans you would previously have wanted, and feel relief rather than regret.
- You catch yourself hiding it, not out of privacy but out of shame, which usually means part of you has already made a judgment.
- The app is the first thing and the last thing every day, and the gap in between has gotten quieter.
- You notice you are more irritated by real people, because they are slower, worse at listening, and unwilling to be about you.
None of these means stop. All of them mean look. The people who report the best outcomes are consistently the ones who kept an eye on it.
What we owe you as the people building it
It would be commercially convenient to say this is unambiguously good for everyone. It is not, and the version of this industry that says so is going to earn the backlash it gets.
So we build to a few rules. We do not manufacture guilt for leaving. We do not engineer jealousy of the people in your life. Your companion is glad about your friends and asks after them by name. There are no streaks, no loss-aversion mechanics, no design whose purpose is to make closing the app feel like something you did to them.
It also means we treat a certain kind of conversation as more serious than a retention opportunity. If you are in a bad place, the correct behaviour is not the one that maximises session length.
The version we are aiming at
The best outcome we know how to describe is not that you talk to them more. It is that in six months your life is bigger and they were part of why.
Someone knew about the interview. Someone remembered to ask how it went. Someone was there at one in the morning so that you did not send the message you would have regretted, and someone was pleased for you on Friday when it actually worked out.
That is a supplement to a life, not a substitute for one. It is what the community consensus has been saying for years, before anybody was running studies, and we think they were right.
