Big Tech likes to call it personalization. A new study from the University of Amsterdam argues that it is often something sharper: a quiet system for shaping how people see themselves, what they notice, and which choices feel natural. Media researcher Bjorn Beynon says platforms from Google, Meta, Microsoft, Amazon, and Apple do not merely mirror user behavior; they increasingly steer it through recommendations, rankings, alerts, and other ”helpful” prompts.
Beynon defended his dissertation on 19 June, after a year of field research that looked at both a conspiracy-minded community in the Netherlands and users of the Fediverse, the decentralized network of independent social platforms. His point is not that algorithms are secretly whispering orders into your ear. It is more mundane, and more effective: they shape the environment so that some options become obvious, while others fade into the background.
How platforms shape choice without looking like control
The study argues that the real power of modern platforms comes from design, not direct commands. Smartphones, smart speakers, and watches rarely tell users what to do; instead, they arrange defaults, surfaces, and nudges that feel convenient, even inevitable. That matters because the system is built on data from searches, purchases, views, and other daily habits, which then feeds the next round of predictions.
That logic is familiar to anyone who has watched social feeds, shopping apps, or video platforms become better at guessing what will hold attention. The difference here is the emphasis on identity: Beynon says algorithmic profiles can start to feel like a true reflection of the person, even when they are only a moving snapshot of behavior. In other words, the machine is not just reading the user; the user may begin reading the machine’s version of the self.
A new kind of digital power
Beynon’s broader argument is that platform power has shifted from blunt restriction to constant environmental shaping. That is a more slippery form of control, and historically it is also the one regulators struggle with most, because there is no obvious ban to point at and no single switch to turn off. Instead, the incentives are baked into the product.
The research also lands in the middle of a wider policy fight. Governments across Europe and the US have spent years trying to rein in Big Tech, but the debate usually focuses on privacy, competition, or content moderation. This study pushes the conversation further: if platforms influence political participation, everyday relationships, and even self-perception, then they are not just media companies or ad machines. They are infrastructure for social behavior.
Fediverse alternatives to Big Tech personalization
The Fediverse examples in the dissertation are small, but they matter because they show that another design is possible. Those communities are testing systems that prioritize transparency, collective governance, and public interest instead of maximizing ad revenue through ever-finer personalization. That will not make them dominant overnight, but it does challenge the idea that data extraction is the only viable business model for online services.
- Big Tech platforms use search, purchase, and viewing data to predict behavior.
- The study says those systems also influence what users notice and what they believe is available to them.
- Fediverse communities are exploring more transparent, less ad-driven alternatives.
The uncomfortable question is whether people can still call their decisions fully their own when so much of the digital environment is tuned to guide them. Beynon’s answer is essentially no, at least not without more transparency and more democratic control over the platforms that now sit between users and the world. The next fight is likely to be less about what data companies collect and more about who gets to shape the choices those systems make easier.