For many of us, the day starts with a swipe. Before even getting out of bed, we might scroll through TikTok, Instagram Reels, or YouTube Shorts. A few clips turn into a few dozen. Minutes easily dissolve into an hour on social media—and later in the day, we return for more.
These apps are deliberately designed to overwhelm the brain's pleasure circuitry and keep us watching. But researchers warn that short-form videos may only be the start when it comes to harvesting human attention.
A recent review by Germany's University of Bayreuth—analyzing 42 studies involving nearly 30,000 mostly adolescent and young adult participants—examined a factor often overlooked in public debate: the mechanics of the platforms themselves.
A "Supercomputer Pointed Directly at Your Brain"
The researchers identified three core features common to short-form video platforms:
Algorithmic Personalization
The Unpredictability of the Infinite Scroll
The Novelty of Rapidly Switching Videos
These features create a media environment entirely unlike traditional television or older social networks. Algorithms continuously select content so users rarely have to make active decisions, and there is effectively no natural stopping point.
"TikTok is fundamentally different from television," Aza Raskin, co-founder of the Center for Humane Technology, told DW. "On the other side of a TikTok screen is a massive supercomputer pointed directly at your brain. It is trained on the behavior of 3 billion other human primates."
Raskin describes the attention economy as a "race to the bottom of the brainstem," noting that if one platform doesn't occupy your time, another will. "This triggers a ruthless knife fight for human attention," he added.
The Dopamine Trap: Hit, Miss, and Desensitization
Scientists have long understood that highly rewarding experiences activate the brain's reward system. Short-form videos (SFVs) are especially effective at exploiting it.
Anna Lembke, a Stanford psychiatry professor and author of Dopamine Nation, explains that moving images are "like catnip for the mammalian brain." SFVs, she notes, are an unusually compelling and potent format, often leading to "video chaining"—a modern equivalent to chain-smoking where it becomes incredibly hard to stop once we start.
Over-stimulating the brain's reward system with unnatural triggers causes a massive flood of dopamine. To protect itself, the brain "downregulates," decreasing the number of dopamine receptors. Repeated hyperstimulation makes the brain less sensitive to reward over time.
"We need more videos over time, not to get high, but just to feel normal," Lembke said. "And we lose our ability to engage in and take joy in other, more modest rewards, like watching a sunset, sharing a meal with friends, or reading a good book."
This endless scrolling leads to reward desensitization. Users need more videos with more extreme content just to get the same baseline effect. The result is a familiar paradox: continuing to search for the next rewarding clip even when the experience itself is no longer particularly enjoyable.
Why Personalization is the Real Engine
While infinite scrolling gets much of the public attention, neuroscientist Ben Rein argues that the real power lies in hyper-individualization.
"Infinite scroll and novelty matter, but they're delivery mechanisms; personalization is the engine," Rein explains.
He describes a typical TikTok feed as a system running thousands of tiny, real-time experiments on you, learning what keeps you watching faster than you do. Personalization becomes incredibly powerful when combined with the unpredictability Lembke describes. The algorithms evolve the "perfect drug of choice" for each unique person, making the exchange feel organic and alive.
This taps directly into our evolutionary wiring. Novelty overcomes our natural boredom, while unpredictability engages the treasure-seeking, foraging parts of our brain—making those rare, highly entertaining clips feel immensely rewarding when we finally swipe to them.
What Comes Next?
Across the 42 studies examined by the University of Bayreuth team, several themes repeatedly emerged among heavy users: higher attentional difficulties, lower working-memory performance, higher anxiety and depression, weaker self-regulation, and addiction-like patterns.
However, lead author Marlene Ebster cautions that the evidence remains limited. "There is still insufficient evidence to conclusively claim that short-form video platforms cause 'brainrot' or extreme dopamine effects," she notes.
Rather than recommending blanket bans, the authors suggest helping young people understand exactly how recommendation systems work and how platform designs manipulate behavior.
But as we learn to navigate today’s algorithms, tomorrow’s are already being built. For Rein, the next frontier is deeper personalization. As AI improves, the gap between what you genuinely enjoy and what you are shown will keep shrinking.
"We are nowhere near the peak of attention engineering," Raskin warns. "The capabilities of generative AI to maximize attention, ranging from synthetic media to highly personalized synthetic relationships, will completely dwarf the power of any psychological engineering we have seen to date."