A new study by researchers at the University of Macau shows that people who are positive about artificial intelligence spend more time on social media platforms – and are more likely to show signs of digital addiction. This finding sheds new light on one of the less obvious factors influencing our relationship with technology: our own beliefs about it.
Algorithm behind the curtain
It is no secret to anyone in the industry that social media today is very much an AI-powered recommendation system. Every scrolled post, every paused second on a video, every click of ‘like’ – these are inputs for models that optimise the content of our feeds in real time. The goal is not to inform or educate, but to keep attention – for as long and as strongly as possible.
In theory, this is excellent optimisation. In practice, designing an environment that can sense a user’s weaknesses and adapt to their emotional profile. And this is where the less-discussed aspect comes in: how do our beliefs towards AI affect the way we use these systems?
The psychology of scrolling – new data
Professor Christian Montag’s team conducted an online survey of more than 1,000 German adults, analysing the relationship between perceptions of AI and intensity of social media use. The results? People declaring a positive attitude towards AI spend noticeably more time on social media than their sceptical counterparts. What’s more – they are more likely to show symptoms indicative of problematic use of these platforms: difficulty detaching themselves, emotional addiction or treating the app as an escape from everyday difficulties.
The correlation was particularly strong among men. In women, differences were also present, but were less pronounced. No similar patterns were observed among those critical of AI.
AI as a trusted companion – and a risky guide
As interpreted by the researchers, it is not just the presence of AI in social media apps that is crucial, but our own attitudes towards the technology. People who see AI as useful, innovative and trustworthy are more willing to cede control to algorithms. Acceptance makes them more willing to follow recommendations and less likely to question what appears on screen.
On the other hand, sceptics are more likely to analyse, question and make informed decisions about what they use and for how long. This does not mean that they are completely immune to engagement mechanisms – but they are more aware of their existence and consequences.
Lessons for digital product developers
For the technology industry, these results are not just a psychological curiosity. It is another signal that designing AI-based products requires more than just optimising KPIs. If enthusiasm for AI increases susceptibility to addictive patterns, then technology companies should take this into account when designing engagement and recommendation mechanisms.
Particularly as people who are positive about AI are often early adopters, opinion leaders and loyal users. Long-term, this can create a tension: how do we build technologies that are effective but do not use the user’s trust as an entry point to overexploit their attention?
In the context of growing discussions about the ethics of AI and the transparency of algorithms, it is worth considering the introduction of balancing mechanisms – such as limits, suggestions for breaks, the ability to review recommendation history or better options for personalising feeds. This is particularly relevant in the context of increasing regulation of recommendation systems in the EU and the US.
Not just algorithms, but beliefs
The Macau study shows that how we perceive technology can be just as important as how it works. Acceptance of AI can open the way to greater trust – but also a loss of control over time and habits. And while AI in social media will stay with us for a long time, it is worth asking ourselves: are we ready to consciously participate in this relationship?
Because if enthusiasm for AI means more scrolling, then maybe it’s time for a new digital competence: critical thinking towards the technology we like best.
