YouTube will remove videos that include “realistic” genAI recreations of people’s faces or voices–if they ask it to

Scarlett Johansson might be glad to hear this one: YouTube now allows people to request that AI-generated content be removed from its platform if the content in question realistically simulates their face or voice.

First spotted by TechCrunch, YouTube’s Privacy Guidelines were updated in June to create a “privacy process for people to request the removal of AI-generated or other synthetic or altered content that simulates an identifiable individual, including their face or voice.” YouTube’s been talking about the change since November 2023, and the above quote is from a post in March 2024 where the platform lays out how it wants creators to disclose genAI/synthetic content.

“If someone has used AI to alter or create synthetic content that looks or sounds like you, you can ask for it to be removed. In order to qualify for removal, the content should depict a realistic altered or synthetic version of your likeness,” YouTube says about the new guideline.

While this is some measure of protection for people, filing a request doesn’t 100% guarantee content will be taken down, and even if it is

taken down, the uploader might not face any penalties, as YouTube’s Privacy Guidelines are separate from its Community Guidelines. A YouTube spokesperson tells Tubefilter that creators are allowed to upload genAI content that simulates real people’s faces and/or voices; it simply reserves the right to take that content down if a privacy complaint is made, and if the video meets certain criteria.

YouTube says it’ll consider “a variety” of criteria when deciding to remove videos, including whether the content “contains parody, satire or other public interest value” and whether it “features a public figure or well-known individual engaging in a sensitive behavior such as criminal activity, violence, or endorsing a product or political candidate.”

It also considers whether the creator of the content disclosed to viewers that their video was made by genAI.

If someone files a privacy violation complaint against a video, the uploader has 48 hours to take action–meaning they can either remove the likeness from their video, or delete the video altogether. If the uploader takes no action within 48 hours, YouTube steps in and initiates a review of the video. If it finds someone’s privacy was violated, it could delete the upload.

And, if the uploader is a repeat offender, they could lose their account, YouTube warns.

“If we remove your video for a privacy violation, do not upload another version featuring the same people. These people will likely file another privacy complaint or report you for harassment,” it says. “We’re serious about protecting our users and suspend accounts that violate people’s privacy.”

YouTube and its parent Google have both been bullish about AI, but at the same time have to contend with the numerous problems that stem from genAI, including concerns about deepfakes. (Though during our conversation, the same YouTube spokesperson tells us there are some people out there who may want deepfakes of themselves on YouTube.) Considering we’re deep into a presidential election year, there are likely to be more guideline updates around AI–on YouTube and elsewhere–in the months leading up to Election Day.