Technology news and discussion on TalkFuse.
What if AI’s biggest impact on music isn’t just what gets made, but what gets wanted?
In an opinion piece for Music Business Worldwide, Deviate Digital CEO Sammy Andrews argues that the music industry's AI debate has focused too narrowly on ownership and payment, while overlooking a more fundamental shift: whether streaming data still reflects genuine audience taste or merely mirrors what recommendation algorithms have already pushed onto listeners. She contends that as discovery moves from browsing platforms to conversational AI assistants, a new and less accountable form of gatekeeping is emerging, one that shapes individual preferences at massive scale while offering little transparency about how its suggestions are formed.
Andrews cites the growing integration of AI into music discovery, such as Spotify's presence within ChatGPT and Google's AI Overviews reducing click-through rates, as evidence that "ask and receive" is replacing "search and click". She describes testing three popular AI tools for music recommendations, finding they gave confident suggestions but vague, untraceable justifications, drawing on blended data from journalism, criticism and fan communities without attributing sources. She warns that chart positions, already influenced by playlist placement, could become further removed from genuine listener demand as the "bottleneck" shifts from algorithmic playlists to AI prompt responses, with platforms holding unused data that could distinguish intentional listening from passive, algorithm-driven plays.
In short
- AI recommendations may reflect platform outputs, not genuine listener taste
- Music discovery is shifting from browsing to AI-driven "ask and receive"
- AI tools struggle to explain or attribute the basis for song suggestions
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
The concern has genuine merit because AI systems increasingly shape musical preferences through opaque mechanisms rather than simply reflecting existing tastes. As discovery migrates from searchable playlists to conversational AI without transparent reasoning, gatekeeping power concentrates in fewer platforms. The data asymmetry—where platforms withhold information distinguishing genuine demand from algorithm-driven consumption—means we lose sight of what audiences actually want rather than what they've been encouraged to consume.
The case against
Modern AI recommendations represent a democratisation compared with historical gatekeeping by radio programmers, MTV editors, and A&R executives, which concentrated power far more severely and offered no transparency whatsoever. These systems serve a genuine need—navigating vast music libraries—and are commercially incentivised to recommend music people actually enjoy rather than manipulate them. Rather than distorting taste, AI often exposes listeners to increased diversity and niche content they wouldn't discover through traditional gatekeeping, whilst remaining more personalised and responsive than prior systems.
Explained simply
Music discovery has moved from people searching for artists and songs to platforms recommending what to hear. Streaming services use listening data to shape playlists and suggestions, so what becomes popular can reflect both listeners’ choices and what the services have put in front of them.
AI assistants are beginning to offer music recommendations in response to questions, drawing on information from sources such as reviews, news and fan discussions. Their explanations may not show clearly which sources shaped a suggestion or how it was chosen.
This matters to artists, listeners and the music industry because recommendations can influence what people hear and what counts as popular. If listening figures include many plays prompted by platforms, they may not show how much people actively sought out the music.
AI Entertainment Music Technology