Infinite scroll solved a product problem brilliantly: users no longer have to decide whether to load another page.
That convenience also removed something easy to overlook—the moment when people could decide whether they wanted to continue at all.
Today, TikTok, Instagram Reels, YouTube Shorts and other feed-based products combine continuous scrolling, autoplay and algorithmic recommendations into an experience with few natural stopping points. New behavioral research suggests those design choices can influence more than session length. They may affect how people experience time, how much control they feel over their behavior and how effectively they direct attention afterward.
For technology companies, that makes infinite scroll more than a user-interface pattern. It is an experiment in human behavior running at enormous scale.
Infinite feeds remove moments for reconsideration
Traditional webpages eventually end. Reaching the bottom creates a decision point: load another page, follow a link or leave.
Infinite scroll removes that interruption by automatically supplying more content.
Research on the ethics of the attention economy describes this as the removal of natural stopping cues that previously gave users an opportunity to reconsider whether they wanted to continue. When new material appears continuously, the interface reduces the number of explicit decisions required to remain engaged.
That distinction matters because friction is not always a product defect.
In many digital products, eliminating unnecessary clicks improves usability. But when the product itself is competing for attention, removing friction can also eliminate moments when users regain conscious control over a session.
Recent studies are beginning to test what happens when those moments are deliberately restored.
Autoplay can change users’ sense of agency
At the 2026 CHI Conference on Human Factors in Computing Systems, researchers Alberto Monge Roffarello and Andrea De Luca experimentally modified TikTok-style feed mechanics to examine how infinite scroll and autoplay affect users’ sense of agency.
Their controlled study found that replacing autoplay with explicit playback significantly increased perceived agency and reduced time distortion. Replacing continuous scrolling with paginated navigation produced less pronounced quantitative effects, but participants nevertheless preferred pagination among the alternatives tested.
The researchers describe infinite scroll and autoplay as Attention-Capture Damaging Patterns, or ACDPs—design mechanisms optimized for engagement that may work against users’ perceived control.
The finding is especially relevant to product teams because the intervention was small. Users did not need to leave the platform or install a digital-wellbeing application. Changing how the feed advanced was enough to alter their experience of control.
It suggests that agency can partly be designed into—or out of—the interaction itself.
Short-video feeds may also distort how time feels
The relationship between feed design and time perception is becoming another area of research.
A 2026 study published in Cyberpsychology, Behavior, and Social Networking followed 151 participants for 14 days using an experience-sampling method designed to observe short-video consumption in everyday conditions.
Rather than finding an immediate change, researchers observed significant differences in participants’ time estimation 15 and 30 minutes after watching endless-loop videos. The study reported a small-to-medium effect of approximately β = 0.20 and emphasized that its sample was predominantly young, educated, female and German-speaking, limiting how broadly the result should be generalized.
That delayed effect is particularly interesting.
People may not simply lose track of time while scrolling. Heavy short-video consumption may temporarily affect temporal orientation even after the feed has been closed.
The researchers did not find lasting changes in emotional valence or arousal beyond brief immediate effects, reinforcing that the time-perception result should not be expanded into claims the study did not test.
The cognitive evidence is broader, but still mostly correlational
Time distortion sits within a larger body of research examining short-form video use and cognition.
A 2025 systematic review and meta-analysis published in Psychological Bulletin combined data from 98,299 participants across 71 studies. Greater short-form video use was associated with poorer overall cognitive outcomes, with a moderate mean correlation of r = -0.34.
The strongest relationships appeared in attention, at r = -0.38, and inhibitory control, at r = -0.41. The analysis found weaker associations with mental-health outcomes, averaging r = -0.21.
These figures require careful interpretation.
Most studies in the literature remain observational. A correlation between heavier short-video consumption and weaker attention does not by itself prove that the platform caused the cognitive difference. People with existing attention or self-control difficulties may also be more attracted to rapid, continuously changing content.
The researchers specifically noted that relatively few studies had examined cognitive domains beyond attention and inhibitory control, leaving substantial gaps around memory and reasoning.
That uncertainty makes experimental studies particularly important.
Experiments suggest attention can affect remembering future tasks
A 2026 Behavioral Sciences paper examined prospective memory—the ability to remember to perform an intended action later.
Researchers first recruited 106 college students and compared participants who watched short-form videos with those watching longer videos. A second experiment involved another 64 students.
They found that prolonged short-video viewing significantly reduced accuracy on both focal and non-focal prospective-memory tasks, while performance on the ongoing task itself was not significantly affected.
When researchers reduced the non-focal memory task to a single cue, the negative effect weakened, suggesting that the observed impairment depended partly on the amount of attention required to keep future intentions active.
That offers a more specific mechanism than the broad claim that short videos “destroy attention.”
Rapid content does not necessarily make people incapable of concentrating. Instead, under some conditions, heavy short-video exposure may consume or reorganize attentional resources in ways that make certain memory tasks harder.
Adding friction helps, but users may dislike it
Designing healthier interfaces creates another problem: people often prefer frictionless products.
A separate CHI 2026 study involving 48 participants added friction to TikTok’s autoplay, infinite-scroll and social-investment mechanics while measuring behavior, eye movement, user experience and perceived agency.
Researchers found that disabling autoplay and increasing scrolling friction reduced compulsive engagement patterns. But the interventions also reduced reported user experience and produced only limited or sometimes negative improvements in agency.
That trade-off matters enormously to builders.
A feature can reduce compulsive behavior while simultaneously making the product feel worse. If teams optimize only for retention, session duration or satisfaction scores, the design that best protects user control may appear objectively inferior.
Product metrics therefore encode values whether teams acknowledge them or not.
The next engagement metric may be user control
None of this means infinite scroll must disappear.
Short-form platforms provide entertainment, information, education and creative distribution at enormous scale. Current research also does not justify claiming that every user experiences cognitive harm.
What the evidence does show is that interface architecture matters.
Autoplay determines whether another piece of content requires an intentional action. Infinite scroll determines whether sessions contain natural endings. Recommendation systems determine whether the next item requires searching. Together, these mechanics shape how frequently users have to actively choose to continue.
For founders and product teams, the harder question may therefore be whether maximizing frictionless engagement should remain the default definition of good UX.
A product can be easy to use while making it harder to stop using.
The next generation of social interfaces could treat stopping cues, explicit playback, session boundaries and visible time signals not as obstacles to engagement, but as tools for preserving user agency.
Infinite scroll succeeded by eliminating the moment when people had to ask whether they wanted more.
The research now suggests that putting some of those moments back may be an important part of designing technology people still feel in control of.
