AI and Creativity in 2026: 15 Billion Images a Day and the Fight Over Who Owns What

An AI-generated image won the Sony World Photography Awards in 2023. The artist, Boris Eldagsen, refused the prize and called it a provocation — proof that the photography world wasn’t prepared for what was coming. Three years later, the provocation has become the new normal. AI generates an estimated 15 billion images per day. Suno and Udio produce music that casual listeners consistently mistake for human compositions. AI-written text is so ubiquitous that distinguishing it from human writing has become effectively impossible without specialised detection tools — which themselves are unreliable.

The creative industries are in the middle of the most disruptive transition since the invention of photography — which, it’s worth remembering, was also predicted to “kill art.” It didn’t. But it fundamentally changed what art meant, who could create it, and how creative professionals earned a living. AI is doing the same thing, faster and at greater scale.

The tools reshaping creative work

In images, Midjourney dominates for quality and aesthetics, while Adobe Firefly has become the professional standard because it’s trained exclusively on licensed content — no copyright lawsuits. Stable Diffusion remains the open-source alternative for developers and hobbyists wanting full control. DALL-E 3, integrated into ChatGPT and Microsoft products, is the most accessible for casual users.

In music, Suno emerged as the breakout product of 2025 — generating full songs with lyrics, vocals, and instrumentation from text prompts with startling quality. Users specify genre, mood, tempo, and lyrical themes; the AI produces radio-ready tracks in under a minute. The technology is impressive enough that several major labels have signed deals with AI music companies for background music licensing — while simultaneously suing them for training on copyrighted songs.

Video generation crossed a new threshold when OpenAI’s Sora began limited rollout in late 2025. Generating photorealistic video clips from text descriptions — previously requiring a full production team — is now available to anyone with a subscription. The implications for advertising and independent filmmaking are enormous. The implications for misinformation and deepfakes are equally significant.

AI creative tools in 2026: key comparison

The copyright war

The legal battles are escalating rapidly. Getty Images sued Stability AI for training on its copyrighted photo library without permission. The Authors Guild sued OpenAI on behalf of thousands of writers whose books were used to train GPT models. Universal Music Group sued Suno and Udio for generating music that mimics specific artists’ styles. Visual artists have filed class-action suits against Midjourney, Stability AI, and DeviantArt.

The core legal question remains unresolved: is training an AI model on copyrighted works “fair use” — transformative enough to be legal — or mass-scale copyright infringement? Courts in different jurisdictions are leaning different ways. The EU AI Act requires transparency about training data. Japan has taken a permissive stance, declaring AI training on copyrighted material generally legal. The US is somewhere in the middle, with cases working through federal courts that could set precedent for years. For creative professionals, the practical impact is real regardless of legal outcomes. Stock photography prices have collapsed as AI-generated alternatives flood the market. Freelance illustration rates have dropped 30–40% since 2023. Music licensing for background tracks — once a reliable income stream for independent musicians — is being undercut by AI alternatives at a fraction of the price.

How professionals are adapting

The creative professionals thriving in 2026 aren’t the ones fighting AI — they’re the ones integrating it. Graphic designers use Midjourney for rapid concept exploration, then refine in Photoshop. Video editors use Runway to generate B-roll and background elements, spending their time on storytelling rather than stock footage searches. Writers use AI for research synthesis and first-draft brainstorming, then apply their voice and expertise to transform generic output into distinctive work.

The skill set that matters has shifted. Technical execution — the ability to draw, photograph, or produce music at a professional level — is less differentiating when AI can produce technically competent work in seconds. What’s more valuable now is creative direction: knowing what to ask for, recognising what’s good, curating and editing AI output, and bringing the human perspective that AI fundamentally lacks. The best creative professionals in 2026 are becoming directors and editors of AI-generated raw material.

The bigger picture

AI won’t replace human creativity. It will — and already is — replacing certain types of creative labour. The distinction matters. A machine can generate an image that looks like art, but it has no intention behind it, no lived experience to draw from, no emotional investment in the result. Human creativity emerges from consciousness, culture, pain, joy, and the desire to communicate something meaningful. AI creativity emerges from statistical patterns in training data. Both have value, but they serve different purposes, and the creative industries are still figuring out where the boundaries lie.

Frequently asked questions

Can AI-generated images be copyrighted?

In the US, the Copyright Office has ruled that purely AI-generated images without human creative authorship are not copyrightable. However, images where a human makes substantial creative choices — selecting, arranging, and modifying AI outputs — may qualify for copyright protection for those human-authored elements. The legal framework is still developing, and different countries are taking different approaches. For commercial use, Adobe Firefly’s licensed training data provides the safest option regardless of copyright outcome.

Is AI music legal to use commercially?

It depends on the platform and intended use. Suno and Udio are currently in litigation with major labels, creating uncertainty about commercial use of their outputs. For commercially safe AI music, platforms like Epidemic Sound’s AI features and Mubert use licensed training data with clearer commercial rights. For background music in YouTube videos, podcasts, or business use, check the specific platform’s terms of service regarding commercial licensing — they vary significantly.