Twitter appears to disable autocomplete in its search bar after users pointed out that the feature was recommending gory videos of animal abuse and war footage
Matthew Gault / VICE :
Context & Ripple Effects
Search autocomplete has been a recurring embarrassment for every major platform: YouTube pulled disturbing autofill results in 2017, Facebook apologized for a search box that surfaced sexual and child-abuse videos the following year, and Google was still serving vile suggestions on gender- and race-related queries more than a year after its first fixes. The difference now is whose platform it happens on.
Twitter disabled autocomplete only after users flagged gory animal-abuse and war footage in its suggestions — a notable move from a company that had adopted deliberately free-speech-maximalist moderation under Musk, and one that lands while X's video features are already drawing complaints about gun violence and police brutality surfacing in feeds.
First-order effects
- Every Twitter user loses search autocomplete entirely — a blunt fix that trades a core discovery feature for eliminating the harmful suggestions, rather than filtering them.
- The incident hands advertisers another data point on content risk at X, which reported $1.08B in Q2 2022 ad revenue before Musk and $4B for 2021, making brand safety directly tied to the platform's main income line.
Second-order effects
- X's moderation posture gets squeezed from both ends: the same period saw users report ads for AI 'undress' apps while TikTok and Meta proactively blocked nudify search terms, leaving Twitter looking reactive rather than preventive on search safety.
- Competitors can now point to autocomplete hygiene as a differentiator — Meta, YouTube, and Google each went through their own scandals (YouTube's 2017 autofill results, Facebook's 2018 apology) and built review processes Twitter's stripped-down team may not match.
Third-order effects
- A decade of the same failure mode — Facebook, YouTube, Google, now Twitter — suggests suggestion systems are structurally hard to moderate, pushing platforms toward disabling or heavily constraining them instead of trusting ranking models.
- If harmful-suggestion incidents keep landing during the free-speech-maximalist era, advertiser pressure rather than internal policy becomes the binding constraint on what X's search and feeds surface.
The trend: Platform search and recommendation surfaces keep producing the same harmful-content scandals across companies and eras, forcing moderation by shutdown rather than by tuning.