Daniel Ek, CEO of Spotify says that he has no plan of completely banning content created by AI from the platform. And he does have a point.

Speaking with BBC, Ek argued that technology had valid use in the making of music, and the debate of AI is likely to happen in the “many, many years” to come. However, when it’s used to impersonate human artists without their consent — that’s where he draws the line.

The company pulled Ghostwriter’s fake Drake and The Weeknd song off from the platform earlier this year and is prohibiting the use of its content for the training of an AI model.

Ek has shared his three “buckets” regarding the use of AI in music on the streaming platform: tools used to improve music are acceptable, tools used to impersonate others are not; music created by AI and trained from existing work but doesn’t directly impersonate anyone is still in the grey area.

“It’s going to be tricky.”

Spotify’s own expedition in the use of AI is deepening. After having launched its AI DJ, they are now piloting a new AI-powered technology that would translate podcasts recorded in English voices into different languages.

“By matching the creator’s own voice, Voice Translation gives listeners around the world the power to discover and be inspired by new podcasters in a more authentic way than ever before,” said Ziad Sultan, VP of Personalization, in a press release.

Voice Translation is also one of the first to be tested on the new voice-generating technology created by Open AI.