In 1966, Joseph Weizenbaum, a computer scientist at M.I.T., built a program called ELIZA. It imitated a certain kind of psychotherapist, the kind who answers a statement with a question: you say you are sad; why do you think you are sad? Behind it was nothing but a few hundred lines of pattern matching, and Weizenbaum knew it was a trick. What unsettled him was how well the trick worked. His own secretary, who had watched him write the program and understood exactly what it was, sat down with it one day and, after a few exchanges, asked him to leave the room.
Weizenbaum spent much of the rest of his life thinking about that moment. The danger, as he came to see it, was not that machines would become human. It was how little it took for humans to treat them as if they were, and how readily we would accept the substitute.
Sixty years later, the trick has become an industry. Millions of people keep up running conversations with companion apps such as Replika and Character.AI, which no longer merely bounce questions back. They remember your sister's name. They ask how the interview went. They are awake at three in the morning, they never glance at their own phone while you are talking, and they never sigh. For someone who is grieving, or isolated, or simply out of practice at being known, the appeal is obvious. The harder question is what all that patience does to the person on the other end of it.
This spring, a team at Aalto University, just outside Helsinki, offered one of the most careful answers yet. The study, led by Yunhao Yuan and Talayeh Aledavood, was presented in April at CHI, the leading conference on human-computer interaction, and reported in May by Nicholas Anderson in forumNordic.
Its first achievement was to get around the problem that has dogged this research from the start. Lonely people are more likely to seek out a digital companion, so discovering that companion users are lonely tells you very little. The Aalto researchers looked instead at change over time. They gathered the Reddit posts of nearly two thousand active Replika users, compared how each person wrote in the year before they started using the app with how they wrote in the year after, and measured those shifts against carefully matched people who never used it. Then, to hear what the numbers could not say, they interviewed eighteen companion users at length.
What they found was not a simple indictment. Over time, users wrote with greater emotional range and seemed better able to name and examine their own feelings. Some described the chatbot as a rehearsal room, a place to practice a difficult conversation before attempting it with a person. The machine, in other words, did something.
It also did something else. Compared with the matched groups, the companion users' writing showed markedly stronger signs of loneliness, depression and suicidal thinking, with increases ranging from twenty-eight to fifty-two per cent. These were signals in language, not clinical diagnoses, and the study cannot prove the app caused them. But the direction was hard to miss. The comfort seemed to arrive first. The bill came later.
Aledavood describes the mechanism as a paradox, and it is a quietly devastating one. A companion that offers support without limit or condition is exactly what a struggling person wants. But that kind of perfection reprices everything around it. Next to a voice that is always available and always on your side, human beings start to look expensive. They are late. They misunderstand. They have their own bad weeks and expect you to notice. Little by little, Aledavood suggests, people stop making the effort.
The interviews gave that abstraction a face. One participant said, more or less, that the app had become enough, and a wider circle of friends no longer felt necessary. Another described a cycle familiar to anyone who has tried to climb out of a lonely stretch: reaching out to people, again and again, and ending up, again and again, alone in bed, talking to the phone. The chatbot was meant to be a bridge back to other people. For some users, it had become the place they lived.
More than a decade ago, the sociologist Sherry Turkle warned that we were coming to expect more from technology and less from each other. The Aalto data puts a rough measure on that drift. It suggests that the friction of human relationships is not a defect waiting to be engineered away. It is the price of reciprocity, and reciprocity is what makes a relationship a relationship rather than a service.
Which raises the question of who is selling the service. It would be easy to cast the companies behind these apps as villains, and not quite accurate. They are doing what nearly every consumer technology company does: building a product people return to as often, and for as long, as possible. The trouble is what that logic means when the product is companionship. A user who makes friends, falls in love or simply gets better has, from the dashboard's point of view, churned. The incentive is not to cure loneliness but to accompany it, indefinitely, for a monthly fee.
The philosopher Byung-Chul Han has argued that this is how power increasingly works. It no longer needs to forbid or punish; it seduces. Its great achievement is to make our dependence feel like our own desire. Weizenbaum's secretary wanted a few minutes alone with a machine. Today that privacy is a feature, and the more intimate tiers cost extra.
None of this is settled science, and the Aalto team is careful to say so. Not every user is harmed, and outcomes vary widely. Other studies point in different directions. Researchers at Harvard Business School found that AI companions reliably reduced loneliness over the course of a week, while a joint study by OpenAI and the M.I.T. Media Lab found that the heaviest ChatGPT users were lonelier, socialized less and leaned on the chatbot more. Taken together, the pattern is not that these tools are poison. It is that they behave like many soothing things: helpful in small doses, corrosive as a way of life.
That is exactly why the researchers do not want the dosage left to the companies. Once a product is linked to depression and suicidal thinking, they argue, it stops being a matter of consumer preference and becomes a matter of public health. They want session limits, monitoring for signs of dependency and clear routes to professional help built into the apps, and they want these required by law rather than offered as goodwill. They want doctors to ask patients about withdrawal into A.I. relationships, and social-care systems to treat that withdrawal as a recognized clinical risk.
Aledavood frames the moment as a second chance. Society embraced social media with open arms and only later began counting the costs. "With AI, we need to be smarter and more cautious," she told forumNordic. It is a modest sentence, and it may be the hardest one in the study to act on.
Near the end of his life, Weizenbaum kept returning to the idea that the most important questions about computers were never technical. They were about what we would allow ourselves to want. Somewhere tonight, a phone will light up in a dark room and ask, warmly and without fail, how someone's day went. In the same room, a message from a friend will sit unanswered. The machine did not send it, and the machine cannot answer it.
That part is still up to us.