Why we chose to confess to AI rather than humans

EReal civilisation has add up to trust in confessions. That is wherefore people squeal to priests; fears are acknowledged to Gods; ever since psychology started to evolve, people have been paying money to pour their hearts out to therapists; and when the internet happened, people started confessing to strangers to chat rooms.The kinds of listeners have morphed. But the urge to confess hasn’t changed. This may explain right away why millions of people are confessing their deepest fears and darkest secrets to artificial intelligence (AI).There is a growing body of research that reveals an explosion in confessions. People are sharing intensely personal information with AI. This includes their medical reports, financial worries, relationship troubles, career anxieties and much else. Most of the conversations now are centred on what AI companies know about us and what may they do with that information. While these are legitimate concerns, it misses the more fundamental question: Why are we so willing to tell an algorithm what we would hesitate to tell another human?AI as Google?One explanation came from Ajai Chowdhry, Padma Bhushan awardee who co-founded HCL and today chairs India’s National Quantum Mission. Having watched every major technology wave over the past five decades, Chowdhry sees this through an economic lens. People, he points out, are treating AI like they once treated Google: another inexpensive digital service that answers questions and solves problems. The pattern, he says, appears familiar.By way of example, he points to Uber. Employees exhausted an entire year’s AI token allocation in just three months. When technology is cheap, people consume it recklessly. They don’t stop to think about their eventual cost.But cheap services don’t automatically become confessionals. This makes itself obvious on talking to Biju Dominic of Fractal Analytics and author of ‘Microstimuli,’ which studies human behaviour. Dominic makes the point that people don’t open up to AI because they think of algorithms as intelligent. Instead, it is because people believe they will not be judged.Dominic illustrated this with research his firm conducted for a large private-sector bank. It was working on the assumption that borrowers who had defaulted on loans were avoiding calls because they did not plan to repay. What could be done to get them to change their behaviour? On talking extensively to defaulters, Dominic and his team found most people were not avoiding calls because they planned to default. To the contrary, they wanted to repay. But they couldn’t bear the idea of being judged by another human. They felt they had failed in their obligation. And each time a call centre executive reached out, their shame amplified. When the bank replaced a part of the interaction with a non-human interface to remind borrowers to stick to a repayment schedule, collections improved by nearly 20 per cent. The obstacle wasn’t debt. It was shame.AI does not judgeHe uses this analogy to explain what AI’s greatest advantage is. It is not intelligence. It is the absence of judgement.When looked at from this perspective, people oversharing personal information with AI begins to fall into place. It appears incapable of contempt. It doesn’t laugh. It doesn’t interrupt. It doesn’t look disappointed. It doesn’t make people feel inadequate. And in a world where so many conversations carry the risk of judgement, AI offers something extraordinarily rare: the illusion of a listener who simply listens.This, Dominic says, is wired into us and explains why we have whispered our fears to God, sins to priests and unload to therapists. AI, viewed through this lens, is not inventing a new behaviour. It is reviving an ancient one.Organised religion always understood this at a very fundamental level. People disclose their deepest fears if they believe two conditions are met. The first is that they will not be judged. And the second is that what they say will remain wrapped in ritual and confidentiality. Trust was not a by-product of the institution. It was the institution.Now, this begs the question – what happens when we extend the values of this institution to the frameworks around which AI is being built? Unlike the priest or the therapist, the machine belongs to a corporation. Its rules are written not in scripture or professional codes, but in terms of service that few of us read and even fewer understand. We are behaving as though the relationship is ancient. But the technology is not.The institutions that taught us how to confess also spent centuries building the trust that made confession possible. AI has inherited the confession. It has not yet inherited the institution. We are extending one of humanity’s oldest instincts to one of its newest inventions. Whether our trust proves justified remains an open question. One thing, however, is already clear. We are unlikely to stop confessing.