TikTok launched a public verification tool Thursday that lets anyone upload a video to check whether it was generated by the platform’s own AI suite, a move unveiled at its Safety and Innovation Forum in London.
The detector, accessible through the Transparency Center, returns a result in seconds and does not require a TikTok account. Users must download the content they want to verify and upload it to the page, which supports common formats including JPEG, PNG and MP4, with a size limit of 20 megabytes per image or 100 megabytes per video or audio file.
The tool scans for two signals embedded by TikTok and its parent company ByteDance: metadata credentials and invisible watermarks. It covers content created or edited with TikTok AI as well as ByteDance applications CapCut, Dreamina, Hypic and Pippit, supporting common formats up to 20 megabytes for images and 100 megabytes for video or audio.
TikTok said the tool deletes uploaded material immediately after verification and never uses it to train models, while cautioning that it cannot detect AI content produced outside its own ecosystem.

Company sources said users have created more than 5 billion videos with TikTok’s AI tools. Although such posts carry labels, some evade them.
“El contenido sobre abuso a menores está totalmente prohibido en TikTok, pero los criminales encuentran la manera. Para subsanar estas lagunas estamos mejorando las medidas de seguridad y protección,” said Valiant Richey, global head of Partnerships, Elections and Marketplace Integrity, addressing journalists at the forum where the Spanish foundation Maldita.es had recently exposed more than 1,000 child sexual abuse videos circulating on the platform.
Alongside the verifier, TikTok introduced an opt-in feature called Likeness Detection. Verified creators who meet eligibility requirements can register by confirming their identity through a trusted third-party partner. When the system spots content or accounts that may be using their likeness without consent, it notifies them so they can review and report the material directly to safety teams. The feature is designed to help public figures, including government members, politicians and party accounts, detect and report AI-generated content that uses their image without permission.
“Aquellos que infrinjan nuestras Normas de la Comunidad sobre representaciones dañinas, engañosas o no consentidas de personas serán eliminados,” said Tom Varghese, global head of AI Public Policy.
The rollout coincides with the European Union’s AI Act transparency rules, which took effect on 2 August and require labelling of deepfakes and other AI-generated content, as Europe faces a wave of AI-driven disinformation and deepfake campaigns affecting electoral processes with nine more European countries heading to the polls in the next 12 months.
TikTok launched internationally in September 2017 after ByteDance spent nearly $1 billion to acquire Musical.ly, merging the two apps on 2 August 2018. By April 2020 the platform had surpassed two billion mobile downloads worldwide. It is owned by Beijing-based ByteDance and operates from Singapore and Los Angeles, though it is blocked in mainland China where the sister app Douyin, launched on 20 September 2016, serves the domestic market instead.
Parental controls were also expanded. Guardians can now see who their teens interact with, follow, block or message — though not the content of those conversations — and receive alerts when a teen posts, including the post’s audience reach. Screen-time data will be more visible.
TikTok said it is working with adolescent mental health experts from Boston Children’s Hospital and the International Association for Suicide Prevention to develop future alerts for parents when minors repeatedly search for self-harm or suicide-related terms, according to Elnacional. A notification will inform parents in the future if minors carry out repeated searches related to self-harm or suicide. The initiative will count on guidance backed by mental health specialists.