SoundCloud Just Updated Their Terms of Service After AI Policy Backlash

You might be having a bad week, but AI is having a worse one. First there was the “racism glitch” that beset Grok, and now music platform SoundCloud is facing some serious criticism for a clause buried in its terms of service.

The imbroglio began in February 2024, when SoundCloud quietly changed its TOS to include:

In the absence of a separate agreement that states otherwise, You explicitly agree that your Content may be used to inform, train, develop or serve as input to artificial intelligence or machine intelligence technologies or services as part of and for providing the services.

That quietly sat there in the TOS for more than a year, but this week, Ed Newton noticed the change and posted about it on his X account. The response was immediate and fiery, with many musicians and SoundCloud users decrying the use of their music to train AI.

But some of that user ire seems misplaced; it’s a more nuanced situation than it seems at first.

What Soundcloud was planning to do with AI

It’s easy to see why musicians wouldn’t want their art used to train machines designed to replace them, but, according to SoundCloud, the TOS change was never about that.

The company’s president, Eliah Seton, issued an open letter on Wednesday explaining that they were using AI for “powering smarter recommendations, search, playlisting, content tagging, and tools that help prevent fraud” but the company has never “used artist content to train AI models. Not for music creation. Not for large language models. Not for anything that tries to mimic or replace your work.”

According Seton, it’s basically been a misunderstanding. “The language in the Terms of Use was too broad and wasn’t clear enough. It created confusion, and that’s on us,” Seton wrote.

Opt-in or opt-out: The eternal question

SoundCloud may have cleared up how it has used AI in the past, but company reps waffled on what it plans to do with your music in the future. The initial response to the controversy, delivered in a statement to Verge from Marni Greenberg, SVP and head of communications at SoundCloud, explained, “Should we ever consider using user content to train generative AI models, we would introduce clear opt-out mechanisms in advance.”

The community responded with, “shouldn’t that read ‘opt-in?'”

“Yeah, opt-in. Sounds great,” SoundCloud responded, eventually.

How SoundCloud is planning to change its terms of service

In his open letter, SoundCloud CEO Seton got specific about planned changes to the service’s terms of service. The offensive AI section will be replaced with:

We will not use Your Content to train generative AI models that aim to replicate or synthesize your voice, music, or likeness without your explicit consent, which must be affirmatively provided through an opt-in mechanism.

So AI won’t be used for replication or synthesis of users’ music unless they opt in. Presumably, SoundCloud’s will continue to use AI for recommendations, tagging, and play-listing, a much more benign, and generally accepted used of the technology.

A couple of lessons from this flare-up: One, dealing with AI requires companies to be crystal clear with their users about how AI will be employed. Two, we should all read the TOS.