SoundCloud updates AI policy after backlash: ‘AI should support artists, not replace them’

Following contention complete a quiet update to nan position of usage that seemed to licence nan contented uploaded to nan streamer being utilized to train AI, SoundCloud has updated its AI policy.  

Futurism had antecedently reported that SoundCloud “quietly” updated its position and conditions successful February 2024 in which users were “explicitly agreeing” by utilizing nan level to person their contented utilized to train AI. 

The argumentation read: “You explicitly work together that your Content whitethorn beryllium utilized to inform, train, develop, aliases service arsenic input to artificial intelligence technologies arsenic portion of and for providing nan services.” 

Users were understandably upset and now, pursuing backlash, SoundCloud CEO Eliah Seton has responded regarding nan platform’s stance connected AI and “how contented whitethorn interact pinch AI technologies wrong SoundCloud’s ain platform”. 

The letter, titled “A Letter from our CEO: Clarifying our Terms of Use”, states: “SoundCloud has ne'er utilized creator contented to train AI models. Not for euphony creation. Not for ample connection models. Not for thing that tries to mimic aliases switch your work.” 

It besides states: “We don’t build generative AI tools, and we don’t let 3rd parties to scrape aliases usage creator contented from SoundCloud to train them either.” 

“Our position is simple: AI should support artists, not switch them.”  

Seton went connected to explicate nan updates to nan Terms Of Use past February were meant to explain really “we whitethorn usage AI internally” to amended SoundCloud for its users, including “powering smarter recommendations, search, playlisting, contented tagging, and devices that thief forestall fraud”. 

Now, SoundCloud has changed its Terms Of Use, and it will only usage AI-training connected contented uploaded to nan level pinch users’ consent. 

The caller argumentation reads: “We will not usage Your Content to train generative AI models that purpose to replicate aliases synthesize your voice, music, aliases likeness without your definitive consent, which must beryllium affirmatively provided done an opt-in mechanism.” 

Today, nan Data (Use and Access) Bill returns to nan UK House of Lords for consideration, successful bid to reside really nan government’s desire to foster a British AI manufacture could let exertion companies to circumvent copyright laws and usage imaginative contented to train their models – each without nan support of nan creators. 

Chi Onwurah, nan chair of nan cross-party committee, has antecedently urged nan authorities to bring guardant nan AI information bill. 

Onwurah told nan Guardian: “It’s perfectly captious that nan authorities shows it is connected nan broadside of group erstwhile it comes to technology, peculiarly erstwhile it comes to nan tech platforms and nan effect exertion is going to person successful their lives.” 

This besides comes astatine a clip erstwhile artists person been speaking retired against companies exploiting copyrighted works and informing against “predatory” usage of AI successful music.  

Earlier this year, much than 200 artists featured connected an open missive submitted by nan Artist Rights Alliance non-profit, calling connected artificial intelligence tech companies, developers, platforms, integer euphony services and platforms to extremity utilizing AI "to infringe upon and devalue nan authorities of quality artists.” 

Amongst those names were Stevie Wonder, Robert Smith, Billie Eilish, Nicki Minaj, R.E.M., Peter Frampton, Jon Batiste, Katy Perry, Sheryl Crow, Smokey Robinson, and nan estates of Bob Marley and Frank Sinatra. 

The letter, while acknowledging nan imaginative possibilities of caller AI technology, addressed immoderate of its threats to quality artistry. Those see utilizing preexisting activity to train AI models - without permissions - successful an effort to switch artists and truthful “substantially dilute nan royalty pools that are paid retired to artists.” 

The missive stated: “Make nary mistake: we judge that, erstwhile utilized responsibly, AI has tremendous imaginable to beforehand quality productivity and successful a mode that enables nan improvement and maturation of caller and breathtaking experiences for euphony fans everywhere. Unfortunately, immoderate platforms and developers are employing AI to sabotage productivity and undermine artists, songwriters, musicians and rightsholders. When utilized irresponsibly, AI poses tremendous threats to our expertise to protect our privacy, our identities, our euphony and our livelihoods.”