A beta that ElevenLabs opened in January made plain that a few dozen seconds of audio were enough to produce a convincing copy of someone's voice, and within about a week it was being misused. Motherboard reported that 4chan users had posted clips of Emma Watson's and Joe Rogan's voices reading racist and homophobic material. The company said it was seeing "an increasing number of voice cloning misuse cases" and set out its response: restricting voice cloning to paid accounts, and providing a way to tell whether a clip had come from its own tools. Founded in 2022, it was a young company at this point. Synthetic speech was not new, but a tool anyone could pick up to clone a particular person's voice from that little material was, at the time, something new.