Journal of Digital Consent Science and Unsolicited Media Ethics
Park, R., Ferreira, T., Osei, C.
Department of Digital Consent Science, University of Dovehill
Applied Media Intrusion Research Unit, Harwick University
Received: 07 February 2025 · Accepted: 07 February 2025
This study examines the psychological and behavioral consequences of autoplay video features — digital media that initiates playback without explicit user action — on consent experience, audio environment disruption, and emotional disruption in public and semi-public contexts. The Autoplay Consent Violation Scale (ACVS) was administered to 264 participants following documented autoplay events. Ninety-four percent of events occurred in contexts where audio was not expected, 78% disrupted an existing ambient environment. The autoplay video does not ask. It has never asked. This is the entire finding, stated in advance of the data for efficiency.
Autoplay video — the automatic initiation of video playback upon page load or scroll position, without explicit user action — has become a standard feature of digital news, social media, and e-commerce platforms. Its design rationale is engagement maximization: users who might not actively choose to watch a video may consume it if it begins without requiring a choice. This rationale treats the user's passive exposure as equivalent to their active engagement, a conflation that raises significant consent questions (Sunstein, 2014). The present study provides the first validated psychological assessment of the autoplay consent experience, treating autoplay initiation not as a neutral feature default but as a designed media delivery mechanism whose behavioral and psychological consequences for users have not been formally measured.
Participants.
Two hundred and sixty-four adults (M age = 29.8, SD = 6.1) who had experienced an autoplay event in the prior 48 hours were recruited via digital sampling. Exclusion criteria included individuals browsing with autoplay disabled (n = 37, excluded as having solved the problem, which removed them from the population experiencing it). IRB protocol DM-2024-0134 was approved.
Instruments.
The Autoplay Consent Violation Scale (ACVS; 18 items, α = .88) measured perceived consent violation, ambient environment disruption, startle response frequency, and what participants described as 'visibly reacting in a coffee shop.' Context data for each event were logged (public, home, work, public transit). A control group browsed equivalent content with static images and reported no instances of their phone suddenly announcing something.
Procedure.
ACVS was administered via experience sampling within 30 minutes of each documented autoplay event over four weeks.
Context of Autoplay Events.
Ninety-four percent of autoplay events occurred in contexts described by participants as audio-unexpected — public spaces, meetings, quiet environments, or situations where device volume was not anticipated to be relevant. Seventy-eight percent disrupted an existing ambient environment.
Consent Violation Scores.
ACVS consent violation subscale scores were significantly elevated in all conditions, t(263) = 14.1, p < .001, d = 1.73. The effect was largest for news website autoplays that began with an advertisement (η² = 0.24).
Social Disruption.
Sixty-one percent of autoplay events produced visible social disruption — participant behavior that was observable to others in their vicinity. Mean self-reported embarrassment following public autoplay was 4.8/7 (SD = 1.2).
The 94% audio-unexpected context rate establishes that autoplay video reliably activates in conditions where the user has not prepared their acoustic environment for media playback. This is not an edge case. It is the modal experience of autoplay: the user is doing something, and then they are also involuntarily broadcasting something.
The social disruption finding (61% visible, mean embarrassment 4.8/7) documents a social cost of autoplay that falls entirely on the user whose device initiated the event without their action. The user is embarrassed. The platform is unaffected. The asymmetry of this outcome is the consent violation's most precise expression.
The news advertisement autoplay — producing the study's largest consent violation effect — represents the specific case where the user arrives for content, is confronted with an audio advertisement before consent, and has no autoplay-free option because the mechanism operates before any user choice can be made. This is, in the authors' view, simply aggressive.
Autoplay video initiates playback in 94% of audio-unexpected contexts, produces social disruption in 61% of events, and generates consent violation scores that are largest when the unsolicited content is an advertisement. It does not serve the user. It serves the platform's engagement metrics at the user's acoustic and social expense. The authors recommend autoplay-off as the universal platform default and acknowledge that this recommendation will not be followed without regulatory intervention.
Correspondence: park@of-dovehill.ac