By Key Points
- YouTube allows individuals to request the removal of AI content impersonating their face or voice.
- This is part of a wider privacy policy update by YouTube.
- Not all requests result in takedowns; each case will be reviewed by YouTube individually.
- The policy focuses on fake content about public figures and sensitive topics.
YouTube now lets users ask for the elimination of videos made with artificial intelligence technology that simulates their likeness or voice. The change is one of several updates YouTube has made over time in response to concerns about privacy and deepfakes on its platform.
Under these rules, people can report videos created using AI under YouTube’s privacy complaint process. This builds on the company’s responsible AI efforts, which were launched in November. Previously, such content could be flagged as misleading — now it can be flagged for removal as a violation of privacy.
Factors Affecting Content Removal
However, making a request does not necessarily mean that the video will be taken down. Each case will go through a review process during which the following factors will be considered:
Disclosure and Identification: Was the content sufficiently labeled as being synthetic or artificially intelligent (AI)-)-based? Does it uniquely identify someone?
YouTube now lets you request removal of AI-generated content that simulates your face or voice https://t.co/1Ey89OjE33
— Sarah Perez is also on Threads, Mastodon & Bluesky (@SarahPerezTC) July 1, 2024
Type Of Content: Is it a parody? Satire? Does it have public interest value?
Public Figures/Sensitive Situations: Does it involve public figures or show individuals engaged in potentially embarrassing conduct like criminal activity or political endorsements?
AI-generated endorsements can be especially problematic in an election year because they could sway voters’ decisions.
Process For Taking Down Content
If somebody files a complaint against your video, you’ll have 48 hours to resolve that concern before YouTube takes action and places strikes against your channel. If you remove the video within that window, then the complaint goes away with it — but if you don’t, then YouTube will step in and review things for itself.
When a video is removed from YouTube, it’s wiped off the site entirely. This includes any personal data or information about the person who complained being scrubbed from the video’s title, description or tags. You can blur out people’s faces in your videos if you want, but just making a video private isn’t enough — someone could still revert it to public status if they found the URL.
Other Tools and Measures
YouTube has introduced several tools to help deal with AI-created content. In March, it launched a Creator Studio tool that lets creators disclose when their videos contain altered or synthetic media, including generative AI.
More recently, YouTube began testing a feature that allows users to add crowdsourced notes providing additional context about a video — for example, indicating whether it’s parody or misleading.
But while these measures are all well and good — and YouTube has experimented with generative AI tools like a comments summarizer and conversational tool for video recommendations — AI-generated content must still follow YouTube’s Community Guidelines.
Labeling something as artificially intelligent doesn’t excuse it from being removed under those rules if it violates them.
Creators’ Implications
Creators should know that privacy complaints are separate from Community Guidelines strikes against their channels. Privacy complaints won’t lead to an immediate strike on an account; however, repeated violations could still result in action being taken by YouTube against the offending party account holder(s).
The updated policy reflects YouTube’s goal of fostering innovation around artificial intelligence while also safeguarding people’s right to privacy. The company continues working on these policies in light of new threats posed by deep fake technology and other forms of misleading content created using AIs such as bots which mimic human speech patterns
What's your reaction?
