Read the blog from YouTube here.
YouTube is expanding its AI-powered “likeness detection” tool to include celebrities and the entertainment industry, aiming to prevent the misuse of public figures’ identities in deepfake content.
This initiative is the latest step in YouTube’s efforts to address AI-generated impersonation, which is an increasing concern as synthetic media becomes more realistic and accessible.
YouTube announced that its likeness-detection technology is now available to celebrities, talent agencies, and management firms, allowing them to identify and address AI-generated videos that imitate their faces. The feature does not require celebrities to maintain their own YouTube channels and can be accessed by representatives acting on their behalf.
Major industry players such as Creative Artists Agency (CAA), United Talent Agency (UTA), and William Morris Endeavor (WME) have supported the rollout and provided feedback on the system.
How likeness detection works: The technology scans uploaded videos for AI-generated content that replicates a person’s face, enabling the identification of deepfakes and other synthetic media. After detection, users can review flagged content and request removal in accordance with YouTube’s privacy policies. Takedowns are not automatic and may depend on context, including whether it is parody or satire. The system uses biometric data like facial data from videos and selfies to identify likeness.
YouTube also highlighted earlier how content removal decisions are taken after “assessing if an individual is uniquely identifiable based on their image or voice, full name, financial information, contact information and other personally identifiable information”.
Likeness detection data not part of AI training: It had also explained that participants must verify “their identity before enrolling in likeness detection“, and this data is used solely for verification—not for training generative AI models. The platform emphasised that this is backed by legal frameworks like the NO FAKES Act, which establishes “a federal right of publicity” and ensures “technology serves, never replaces, human creativity”.
Expansion builds on earlier rollouts: The celebrity rollout builds on a phased expansion of YouTube’s likeness-detection system, which has progressively widened access to groups deemed vulnerable to impersonation.
In these earlier phases, participants had to enrol and verify their identity, after which they could monitor YouTube for AI-generated videos using their likeness and request removals under the platform’s privacy rules.
YouTube has described the system as analogous to Content ID, but for faces—allowing individuals to detect, review, and flag deepfakes at scale rather than relying on manual reporting.
Gaps and privacy concerns around likeness detection: A key concern is that the tool effectively ties identity protection to platform participation. Earlier versions were limited to YouTube creators, meaning individuals had to join the platform’s ecosystem to safeguard their likenesses, creating a dependency similar to the platform’s copyright enforcement systems. At the same time, the system raises significant biometric and consent issues: users must submit government-issued ID and facial data to generate “face templates” for detection, prompting concerns about the expansion of sensitive data collection without parallel protections in copyright law.
There is also ambiguity around how the tool intersects with fair use and parody, as AI-generated likenesses can be challenged under privacy rules rather than copyright law, potentially enabling takedowns of content that might otherwise qualify as legitimate expression. Further uncertainty surrounds the scope of likeness rights, particularly for deceased individuals or in regions with underdeveloped legal frameworks—such as India—where gaps in privacy and personality rights enforcement could limit the system’s effectiveness.
Also read:
For You
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。