By Calli Tzani, Reader in Psychology, Cyberpsychology Research Network Co-Director and ADM Forensic Editor, University of Huddersfield, and Maria Ioannou, Professor of Psychology, Cyberpsychology Research Network Co-Director, MSc Investigative Psychology Director, University of Huddersfield
Social media doesn’t necessarily reflect who you are, but it may intensify particular personality traits and confirm or even persuade you that your worst behaviour is justified.
We are addressing that imbalance, with our framework, the Algorithmic Trait Amplification, or ATA, with particular regard to narcissism.
Narcissism runs on a spectrum. At the far end is narcissistic personality disorder, a clinical diagnosis. But many people have milder narcissistic traits – a strong need for admiration, a habit of overrating their own importance, struggling to take criticism, little empathy. That said, some traits such as healthy ambition and assertiveness are beneficial.
People who have these traits, which are common, are especially drawn to social media and platforms built around status and approval from others.
For years we have been studying online behaviour and the affect of social media upon relationships; including how psychopathy shows up in romantic relationships and how TikTok shapes young women’s attraction to dangerous partners. That work, combined with what psychology already knows on how narcissistic people operate online, how algorithms chase engagement, and what personality change requires led us to a realisation. The algorithmic loop process may slowly strengthen an inflated, entitled self-image.
These are the five steps over which ATA predicts that social media will heighten narcissistic traits.
Stage 1: The engagement hook
Someone with narcissistic traits tends to react strongly to content that matches their self-image. Winning. Status. Being special. Always being right. Most of us have some version of a For You content feed. But for people high in these traits, the pull is more powerful. Such content fulfils a narcissist’s higher need for validation.
Stage 2: Taking it all in
Social media recommendation systems are designed to keep us engaged but they cannot tell whether the content is good for us. They only measure whether it holds our attention. Our reactions become instructions: give this person more of this, feed the loop.
For example, a user rewatches a video about how “strong personalities intimidate weak people”. The app notices and fills the feed with content about cutting off “toxic” friends and refusing to apologise. Every video tells them the problem is everyone else. And they agree. So, they watch more.
Audit studies of online recommendation systems show that feeds narrow in proportion to how selectively a user engages. Someone low in narcissistic traits engages more evenly in different types of content. Someone high in these traits engages hard and narrow and the algorithm obliges.
Stage 3: The mirror oracle
We call this the echo algorithm, after the Greek myth of Narcissus. The nymph Echo can only repeat what is said to her. She cannot challenge or answer back.
Social media algorithms are the same. Over time, the user unknowingly builds a mirror: more praise, more validation, more self-image inflation. And the worst part for narcissistic users is that the feed feels like the world agrees with them. Objective. External. Almost all-knowing. But it is an oracle that they trained themselves, to say exactly what they wanted to hear.
Stage 4: Challenge disappears
Psychologists explain that personality change, in all of us, depends on corrective emotional experiences. These are moments that contradict your self-image in a way you cannot ignore. Or as we call them, uncomfortable moments of truth. They are also how people with narcissistic traits change. These moments help us grow. Take them away, and the traits they would have challenged only get stronger.
Content that challenges the narcissist gets less engagement than content that flatters them. Therefore, challenging content and moments of truths slowly vanishes from the feed – an unplanned algorithmic side-effect of platforms’ aim to chase users’ attention.
Stage 5: The loop feeds itself
Fed a diet of constant validation, narcissistic traits may intensify. Stronger traits produce stronger reactions, thus producing an ever more perfect mirror within the loop.
Is narcissism rising?
Psychologists don’t agree on whether overall levels of narcissism are on the rise. Some psychologists say it is, while others have found no change. So the argument is not settled. But psychology researchers do agree that narcissistic traits correlate with heavier, more problematic social media use and that the expression of narcissistic behaviour is more visible online. Rising or not, our model predicts that for people who already have these traits, the algorithm may be turning up the volume.
Working with narcissistic clients is already considered one of the hardest jobs in psychology. These clients often experience honest feedback as an attack. On one hand, progress depends on carefully managed moments of challenge, built up over sessions. And many clients never get that far: around two-thirds of people diagnosed with narcissistic personality disorder drop out of therapy early. Change is possible, but only for those who stay.
On the other, their phone, trained on years of their own reactions, may offer a world that reflects them at their best, agrees with their grievances and doesn’t challenge. Their therapist may only have an hour a week with them, while the algorithm gets anytime the person picks up their phone.
Our framework may explain why so many people believe narcissism is on the rise: the behaviour is more visible, because the loop rewards it. The algorithm has been running its own experiment on all of us and we wonder, what else has it been intensifying?
