Skip to content
Submit incident
AII-1000-76791
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

TikTok scales back AI video summaries after generating bizarre errors

TikTok scaled back an AI feature that generated video summaries after it produced wildly inaccurate descriptions, including describing a dancer as blueberries. The company modified the tool to only suggest products similar to those shown in videos rather than providing general summaries.

April 2026 · TikTok · Reliability and hallucination
April 2026
AII-1000-767913 sources ↓
Quick facts
Organization
TikTok (subject)
Category
Reliability and hallucination
Country
United States
Source
BBC
Date
April 2026
Drafted by
Raidu Incident Detection

What happened

TikTok's AI Overviews feature, which generated summaries of video content appearing beneath videos, produced wildly inaccurate descriptions in some cases. Examples included describing a video of dancer Charli D'Amelio as a 'collection of various blueberries with different toppings' and similarly bizarre summaries for videos of other celebrities and artists including Shakira and Olivia Rodrigo.

Reported impact

The incorrect and bizarre AI-generated summaries were shared widely on social media and became subject to mockery from users and creators.

Index analysis

The feature was rolled out to some users in the US and the Philippines. Posts reacting to the summaries first began appearing in January, with more examples highlighted in late April.

Governance controls

TikTok scaled back the feature, tweaking its experimental summaries to only suggest products similar to those shown in videos rather than providing broader video summaries.

Classification

Organization
TikTok (subject)
Category
Reliability and hallucination
Industry
Media, social platforms and marketing
Country
United States
Harm status
Potential harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Reputational
Harmed
Consumers
Failure modes
Hallucination, Misidentification
AI systems
TikTok AI Overviews

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

BBC
TikTok scales back AI-generated video overviews after absurd errors
8 May 2026
BBC
TikTok rows back on AI video overviews in US after absurd errors
8 May 2026
Business Insider
TikTok pulls back on an AI feature that described Charli D'Amelio as a collection of blueberries
6 May 2026
Feb 2025TikTok AI blocks 18 Malaysian media accounts reporting mosque assaultConfirmedReliability and hallucinationMARKETING MagazineBERNAMACNA+3Oct 2026AI-Generated Deepfake Videos Spread Disinformation About Ukrainian Military SurrenderConfirmedMisinformation and synthetic mediaУкраинская сеть новостейunn.uaУКРІНФОРМ+5Oct 2026AI-generated deepfake videos spread on TikTok in Russian disinformation campaign targeting UkraineConfirmedMisinformation and synthetic mediaУКРІНФОРМВисокий ЗамокSep 2026Japanese voice actor sues TikTok over AI voice cloningAllegedIntellectual property and appropriationThe Straits Timesnews.hsw.cnSep 2026AI-powered influence campaigns manipulate social media in US and IsraelConfirmedMisinformation and synthetic mediaAnadolu AjansıThe Jerusalem Post