2025 Moved Faster Than Anyone Predicted: Thoughts on AI, Authenticity and What’s Worth Protecting in 2026

As I look back at the year and what a year it has been, one topic has kept rearing its head with increasing urgency: AI and what it means for music. Every conversation, every panel discussion, every quiet pint with someone in the industry has circled back to the same question. Where do we go into 2026? What does the landscape even look like anymore? And more pressingly, how do we navigate a world where the fundamentals of music creation, distribution and discovery are being rewritten in real time?

If you’d asked me twelve months ago where the music industry would be at the end of 2025, I’d have struggled to predict the sheer velocity of change we’ve witnessed. This year moved faster than any I can remember, and the ground beneath the industry has shifted so dramatically that looking back feels like peering into a different era entirely.

Consider where we started. Major record labels spent the first half of 2025 locked in bitter legal battles with AI music platforms Suno and Udio, crying copyright theft and demanding these companies be shut down for training their models on protected catalogues without permission or payment. By late 2025, Warner had settled with Suno and signed what they called a “landmark” licensing deal. What was supposedly an existential threat became, within months, a business partnership. The speed of that pivot tells you everything about where power actually lies in this transition.

Social media platforms tightened copyright enforcement aggressively throughout 2025. TikTok updated its Commercial Music Library terms in July 2025, effectively banning business accounts from using trending sounds and requiring brands to stick to pre-approved tracks. Meta followed suit with stricter screening for sponsored content on Instagram and Facebook. The template for this enforcement had already been set: back in 2022, Bang Energy learned this the expensive way when all three majors sued them for using unlicensed music in over 100 promotional videos, with a federal judge ruling against them. The message was clear: the free-for-all era of social media music usage was over.

Then there was Bandcamp. The platform that had become a genuine lifeline for independent artists changed hands again. After Epic Games bought it in 2022, they sold it to Songtradr in September 2023. The promised artist-first approach dissolved into mass layoffs. Half the staff were let go in October 2023, gutting the editorial team and leaving the community reeling. What had been a carefully cultivated ecosystem was stripped for parts in months.

Against that backdrop of legal whiplash, enforcement crackdowns and platform upheaval, the AI music revolution didn’t pause for breath. It accelerated. And now, as we close out 2025 and look towards 2026, the question isn’t whether AI will reshape music. It’s where artists, tastemakers and genuine music lovers go from here.

Let me be clear about the scale we’re discussing. The global AI in music market hit $5.2 billion in 2024 and is projected to reach $60.44 billion by 2034, with a growth rate of 27.8% annually. Deezer now receives over 50,000 fully AI-generated tracks every single day, accounting for 34% of their total daily uploads. Spotify has deleted more than 75 million spammy AI tracks over the past 12 months alone. These aren’t experimental uploads from curious amateurs, these are calculated attempts to game the royalty system, to flood the ecosystem with content that costs virtually nothing to produce.

Here’s where it gets uncomfortable, and this is personal. I spent time with Suno myself recently. Within minutes, I had a track that sounded disturbingly professional. Not just a melody, a full production complete with stems, mastered to streaming standards, ready to upload. The quality was genuinely unsettling.

Look, I’ve got nothing against Suno as a tool. I can absolutely see legitimate uses for it. Someone making a short film who needs a specific mood cue and can’t afford to license proper music. A content creator needing background audio for a video project. A songwriter sketching ideas quickly to see what resonates. These are all reasonable applications of the technology. It’s a service that helps people make music they need in specific circumstances.

But here’s the problem that keeps me up at night: when someone brand new to music can spend five minutes generating a track, upload it to streaming platforms, and potentially start charting alongside artists who’ve spent years honing their craft, we’ve got a serious issue. Not just for the artists being displaced, but for what it means about value, effort and artistry in music culture.

The traditional pathway into music, however broken and unfair it often was, involved some level of apprenticeship. Learning an instrument. Understanding song structure. Developing an ear. Making terrible music for years before making good music. Building a following. That process, for all its flaws, created artists with knowledge, perspective and something to say. It filtered, at least to some degree, for commitment and genuine creative drive.

Now that barrier is effectively zero. And whilst democratisation of music creation sounds wonderful in theory, in practice it means the ecosystem gets flooded with content from people who’ve skipped every stage of artistic development. They haven’t learned to listen critically. They haven’t studied what makes a track work. They haven’t engaged with music history or culture. They’ve just typed a prompt and uploaded the result.

A survey commissioned by Deezer across 9,000 people in eight countries found that 97% of listeners couldn’t distinguish between fully AI-generated music and human compositions in a blind test. 97%. When quality becomes indistinguishable from algorithmic output, when a five-minute generated track can compete commercially with something an artist laboured over for months, the entire value system collapses.

The adoption rates among actual musicians tell their own story. About 60% of musicians are already using AI tools in some capacity. Electronic music producers lead with 54% adoption, hip-hop follows at 53%. This isn’t a fringe experiment anymore. And I’m not here to moralise about artists using these tools. If AI helps someone realise their creative vision, if it augments their process whilst they remain the creative driver, that’s a different conversation entirely to someone generating content wholesale and presenting it as their own work.

The economic implications deserve attention. A study commissioned by SACEM and GEMA predicts that nearly 25% of creators’ revenues could be at risk by 2028, potentially amounting to €4 billion in lost income. France and Germany alone might see musicians lose €950 million by 2028. Session musicians, sound engineers, video creators who used to make content for labels, writers who crafted press releases, all facing significant disruption. Not because the technology is inherently evil, but because the economics favour replacement over augmentation.

But here’s what interests me more than the doom spiral: the opportunity this creates for those who give a damn about curation. Because if 34% of daily uploads to Deezer are AI-generated now, and Deezer reckons up to 70% of streams for these tracks are fraudulent, we’re about to see the value of human tastemakers skyrocket.

Think about the role of DJs who aren’t chasing commercial success, who spend hours digging through releases to find something genuinely interesting. The selector who prioritises discovery over algorithms. The radio host who introduces listeners to artists they’d never encounter through a recommendation engine. These people become absolutely crucial in an environment where authenticity is invisible and provenance is uncertain.

Except there’s a problem. Proper radio DJs who actually host shows in the electronic music world have largely vanished. Not the ones who submit pre-recorded mixes, but the presenters who talk between tracks, who provide context. What we’ve got instead is someone in a bathrobe eating cornflakes on TikTok calling themselves a tastemaker because they’ve got 50,000 followers. The format doesn’t allow for deep engagement. It’s all surface level, optimised for virality rather than substance.

This is where independent music media becomes more vital. Publications, radio shows, podcasts, club nights, Substack newsletters, community platforms, there’s a whole ecosystem of people doing this work (Decoded included, and thank you if you’re reading this). Not as gatekeepers in the old sense, but as quality filters in an ocean of content where quality has become nearly impossible to assess at scale. The streaming platforms optimise for engagement and retention, not for artistry or originality. Someone needs to do that work, and it’s going to fall to anyone willing to put in the hours.

Research from Sonarworks suggests their base scenario has AI-generated content overtaking human-generated content by 2032. Their fast scenario puts that crossover at 2030. If that projection holds, the people who can navigate that landscape and surface genuine human artistry become invaluable. Consider what’s happening with tagging and transparency. Spotify has introduced a three-pronged policy framework targeting AI-related fraud: cracking down on impersonation violations, implementing spam filters, and working with DDEX to develop industry-standard AI disclosures in music credits. Deezer has gone further, becoming the only major streaming platform to explicitly tag 100% AI-generated content and automatically removing it from algorithmic recommendations.

Survey data shows that 80% of people believe 100% AI-generated music should be clearly labelled, and 73% of streaming users want to know if their platform is recommending AI tracks. There’s genuine appetite for transparency. That appetite creates space for curators who can verify provenance, who know the artists they’re championing, who’ve actually met the people making the music. The challenge for underground artists trying to stand out in this environment is significant. When 82% of listeners can’t tell AI from human music anymore, how does an artist signal authenticity? The answer lies in community and context. Artists releasing through respected labels, featured by trusted tastemakers, playing live shows, engaging directly with audiences, these markers of authenticity matter more now. The streaming platforms benefit from AI-generated content through lower royalty costs, but they can’t manufacture the cultural capital that comes from genuine human validation.

There are AI music magazines now, scraping content from the internet, generating articles with no human oversight. They’re multiplying. The infrastructure of credibility is under pressure. But that pressure creates opportunity for anyone doing the work properly, actually listening, providing context that matters beyond metrics and engagement statistics.

For DJs and selectors, particularly those working outside commercial constraints, this moment offers a chance to reclaim influence. The algorithm can’t replicate the experience of a DJ who understands room dynamics, who knows when to educate an audience and when to give them what they want, who builds a set with narrative and intention. The algorithm can’t replicate the thrill of discovery, of hearing something genuinely new and knowing you’re among the first to recognise its value.

Research suggests live shows could be the biggest winner in this transition, becoming the main channel for human interaction with artists. That makes perfect sense. You can generate a track algorithmically, but you can’t algorithmically generate the experience of being in a room with other people, responding to music together, watching a human perform.

So what does this mean practically? It means building parallel systems. Supporting platforms that prioritise transparency, like Deezer’s tagging approach. Championing artists who are willing to verify their process, to document their creative work, to engage directly with their audience. Creating spaces, both physical and digital, where discovery happens through human curation rather than algorithmic recommendation. It means recognising that the streaming economy, which was already challenging for most artists, is about to get significantly more difficult. About 65% of survey respondents believe it shouldn’t be allowed to use copyrighted material to train AI models, and 73% think it’s unethical for AI companies to use such material without clear approval. Public sentiment is on the side of artists, which creates political and legal leverage. But sentiment takes time to translate into policy, and artists need revenue now.

The answer might be diversification. Direct-to-fan platforms, Bandcamp when it’s not being gutted by new ownership. Vinyl and physical releases for dedicated audiences. Merchandise. Live performance. Building communities that value what you do enough to support it directly. The algorithmic platforms can be part of the strategy, but relying on them exclusively is increasingly risky.

For tastemakers, the path forward involves doubling down on what algorithms can’t replicate: context, expertise, genuine enthusiasm, human connection. Writing that actually engages with music rather than regurgitating press release copy. Radio shows and podcasts that educate and inspire. Club nights that prioritise the music rather than chasing the biggest names. Creating environments where artists can develop without needing to generate algorithmic engagement from day one.

There’s also an opportunity for new models of verification and certification. Human-curated labels and collectives that explicitly signal “this was made by actual people who put in the work.” Quality standards that require documentation of process. Platforms that privilege verified human creators in their algorithms and editorial features. These systems will emerge because there’s demand for them. The survey data proves it.

By the end of 2026, we’re likely looking at an industry that’s unrecognisable from what we had in 2024. The proportion of AI-generated content will continue to rise, potentially hitting 40-50% of total uploads across major platforms. The legal frameworks will still be evolving. More platforms will implement tagging systems. Revenue for human creators will be under pressure. But within that landscape, there’s space for those willing to do the work of curation and discovery properly. The technology will get better, absolutely. We might see the first truly massive AI-generated hit, a track that dominates charts globally with no human performer attached to it. When that happens, it’ll clarify things. It’ll force people to decide what they actually value in music.

The wider issue here isn’t one that any single publication or platform can solve. It’s about collectively recognising that human creativity has specific value worth protecting. That the process of creating art, the years of development and learning, actually matters. That knowledge and context matter. That proper curation requires more than viral moments and engagement metrics. Whether that happens through established publications, new platforms, community radio, club nights, individual tastemakers or some combination of all of these, the work needs doing. And it needs more people willing to do it.

As someone who’s spent years documenting underground music culture, I’m not pessimistic about this moment. Challenged, certainly. Concerned about the people losing work, absolutely. Frustrated that we’ve replaced knowledgeable voices with algorithm-optimised content creators, definitely. But not pessimistic. Because every previous technological disruption in music, from recorded sound to multitrack recording to sampling to digital production to streaming, created new opportunities alongside the disruption. The winners were usually those who adapted quickly, who understood the new landscape and built systems suited to it. The ones who thrived were those who provided value that the new technology couldn’t replicate.

So here’s what needs focusing on, from anyone who cares about this. Championing artists who make music that resonates on a human level, who’ve put in the years, who have something genuine to express. Supporting DJs, radio hosts and selectors who prioritise discovery and genuine knowledge over viral moments. Creating content that adds context and insight rather than just chasing algorithmic trends. Building communities that value authenticity and artistry.

Making it explicitly clear when covering human artists who’ve earned their place through dedication and craft. Not apologetically, but proudly. “This is music made by people who gave a damn, who put time and soul into it, who are worth your attention.” That’s not gatekeeping, it’s navigation. Advocating for transparency and proper labelling. Supporting platforms that take this seriously. Calling out the algorithmic slop when we see it, not with anger but with clarity. “This is what makes music meaningful, and this other thing doesn’t meet that standard.”

Recognising that live performance and physical community are going to become more important, not less. Covering those experiences, documenting them, celebrating them. Creating opportunities for artists to connect with audiences directly. Because that’s where the irreplaceable value lies.

And continuing to do the basic work of music journalism properly. Actually listening. Actually engaging. Actually giving a damn about the artists and the music. Providing depth and context rather than just hot takes and viral moments. That sounds obvious, but in an environment where AI can generate competent copy instantly and influencers can accumulate massive followings without genuine expertise, doing the work with care and attention becomes a competitive advantage. The question isn’t whether AI will be part of music creation. It will be. The question is how we preserve space for human artistry to thrive alongside it, how we help audiences navigate the landscape, how we support the creators who are making music for reasons beyond gaming the algorithm. How we ensure that someone spending years learning their craft isn’t competing on equal footing with someone spending five minutes generating content.

We’re not going back. The tools exist, they’re improving, they’re not going away. But we’re not powerless either. There’s work to do, systems to build, communities to strengthen. The artists making meaningful music will need platforms. The audiences wanting authentic experiences will need guides. The industry will need people willing to draw distinctions and make arguments for quality and genuine creative effort.

That’s where the energy needs to go. Not into resisting the inevitable, but into shaping what comes next. Building the infrastructure that helps human creativity survive and thrive in an increasingly algorithmic world. Because if that work doesn’t get done, if those systems don’t get created, if we don’t dig deeper than the surface level that dominates social media, then yeah, we’ll end up in a bland, algorithmically optimised musical landscape where artistry is an antiquated concept.

But we’re not there yet.

Related