YouTube rolls out new AI deepfake detection tools

YouTube is developing a new process to detect AI-generated deepfakes, to prevent misrepresentation and misinformation on the platform. This latest detection process will notify creators and publishers, if their faces or voice is used in another clip, helping them manage unauthorised content. With the advancement in generative AI, deepfakes have become notoriously concerning, affecting artists and politicians by depicting them in false scenarios.

The company has created 'synthetic-singing-identification technology,' which will allow creators to auto-detect AI-generated content that imitates their singing voices. The technology uses audio matching to identify likely fakes and copies, enabling artists and publishers to manage false depictions more effectively.

In an official blog post, YouTube said, “As AI evolves, we believe it should enhance human creativity, not replace it. We’re committed to working with our partners to ensure future advancements amplify their voices, and we’ll continue to develop guardrails to address concerns and achieve our common goals. Since our earliest days, we've focused on empowering creators and businesses to build thriving communities on YouTube, and our focus remains on fostering an environment where responsible innovation flourishes.” 

Music industry professionals, who already deployed teams to police copyright violations, will benefit from this new technology, adding another tool to their content protection arsenal.

The company is also developing tools to detect and manage AI-generated content depicting real people’s faces, which will be useful for talent agents and celebrities to combat illegal usage. This new role is also expected to benefit political parties; it will help them manage the unauthorised use of political figures' likenesses. These new capabilities will expand the company's existing copyright protection tools, which have already processed billions of claims and generated significant revenue for creators.

Its content ID system has provided rights holders with control over their entire catalogue since 2007, and YouTube is committed to extending this protection into the AI era. The company acknowledges that copyright strikes have become increasingly restrictive over time, but this has given rights holders more control over managing their clients' likenesses. The enhanced protection tools will help YouTube strengthen its relationship with the publishing industry by providing better control over content management.