YouTube might make an official way to create AI Drake fakes

YouTube is currently developing an AI-powered tool that allows users to replicate the voice of famous musicians while recording audio, according to a new report by Bloomberg. The video streaming giant has reportedly approached music companies to obtain the rights to train its new AI tool on songs from their music catalogs. No deals have yet been signed by any major record label, but Bloomberg’s sources claim that discussions between parties are currently ongoing.

YouTube unveiled several new AI-powered tools for creators last month, including AI-generated photo and video backgrounds and video topic suggestions. According to Bloomberg’s report, YouTube had hoped to include its new audio cloning tool among those announcements but was unable to secure the required rights in time.

AI-generated music (much like other variants of generative AI) currently sits in something of a legal gray area due to the difficulties in establishing ownership rights over songs that replicate an artist’s unique voice but don’t directly feature protected lyrics or audio recordings. It isn’t currently clear if training AI voice cloning tools on a record label’s music catalog amounts to copyright infringement, but that hasn’t soured interest in developing AI-generated music features: Meta, Google, and Stability AI have all released tools for creating AI-generated music this year.