AI-Generated “Slop” in Online Biomedical Science Educational Videos: Mixed Methods Study of Prevalence, Characteristics, and Hazards to Learners and Teachers
Bibliographic record
Abstract
Background: Video-sharing sites such as YouTube (Google) and TikTok (ByteDance) have become indispensable resources for learners and educators. The recent growth in generative artificial intelligence (AI) tools, however, has resulted in low-quality, AI-generated material (commonly called "slop") cluttering these platforms and competing with authoritative educational materials. The extent to which slop has polluted science education video content is unknown, as are the specific hazards to learning from purportedly educational videos made by AI without the use of human discretion. Objective: This study aimed to advance a formal definition of slop (based on the recent theoretical construct of "careless speech"), to identify its qualitative characteristics that may be problematic for learners, and to gauge its prevalence among preclinical biomedical science (medical biochemistry and cell biology) videos on YouTube and TikTok. We also examined whether any quantitative features of video metadata correlate with the presence of slop. Methods: An automated search of publicly available YouTube and TikTok videos related to 10 search terms was conducted in February and March 2025. After exclusion of duplicates, off-topic, and non-English results, videos were screened, and those suggestive of AI were flagged. The flagged videos were subject to a 2-stage qualitative content analysis to identify and code problematic features before an assignment of "slop" was made. Quantitative viewership data on all videos in the study were scraped using automated tools and compared between slop videos and the overall population. Results: We define "slop" according to the degree of human care in production. Of 1082 videos screened (814 YouTube, 268 TikTok), 57 (5.3%) were deemed probably AI-generated and low-quality. From qualitative analysis of these and 6 additional AI-generated videos, we identified 16 codes for problematic aspects of the videos as related to their format or contents. These codes were then mapped to the 7 characteristics of careless speech identified earlier. Analysis of view, like, and comment rates revealed no significant difference between slop videos and the overall population. Conclusions: We find slop to be not especially prevalent on YouTube and TikTok at this time. These videos have comparable viewership statistics to the overall population, although the small dataset suggests this finding should be interpreted with caution. From the slop videos that were identified, several features inconsistent with best practices in multimedia instruction were defined. Our findings should inform learners seeking to avoid low-quality material on video-sharing sites and suggest pitfalls for instructors to avoid when making high-quality educational materials with generative AI.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.037 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".