AI-Generated Images and Photographer Copyright: The Growing Legal Conflict
The emergence of generative artificial intelligence has created an unprecedented tension in the creative world. As AI systems become increasingly sophisticated, photographers and visual creators face a fundamental threat to their intellectual property rights. This conflict centres on one critical issue: millions of copyrighted photographs are being used to train AI models without permission, compensation, or even the knowledge of the original photographers.
The Core Problem: Unauthorized Training Data
.The controversy stems from how AI image generators work. Companies developing tools like Stable Diffusion, Midjourney, and other generative AI systems need vast datasets to train their models—often billions of images scraped from the internet. These datasets frequently include copyrighted photographs, stock images, and professional work taken directly from photographers’ portfolios, social media profiles, and image licensing platforms without consent.​
As published by Forbes, Getty Images, one of the world’s largest image licensing agencies, was among the first to take legal action. The company alleged that Stability AI unlawfully copied and processed more than 12 million copyrighted images from its platform to train Stable Diffusion, seeking £1.7 billion in damages. This case was not an isolated incident—by late 2024, over 30 significant copyright lawsuits had been filed against generative AI companies for image copyright violations.
​Beyond Getty’s action, a landmark class-action lawsuit brought by artists Sarah Andersen, Kelly McKernan, and Karla Ortiz against Stability AI, Midjourney, and DeviantArt challenged whether AI companies could legally train systems using copyrighted artwork without permission. These cases have highlighted how the scale of unauthorised use is unprecedented, with AI systems producing approximately 15 billion synthetic images in just one and a half years—an output that has taken photographers over 150 years to accumulate.(Source: The NYU Journal of Intellectual Property and Entertainment Law).
Recent Legal Developments
In a landmark judgement handed down in November 2025, the UK High Court ruled on the Getty Images v Stability AI case. The decision contained significant findings but ultimately favoured Stability AI on the primary copyright claims. Getty had abandoned its main copyright and database rights infringement claims relating to Stable Diffusion’s training during trial, shifting the focus to secondary claims.
However, the court did find limited trademark infringement, determining that some early versions of Stable Diffusion generated outputs containing Getty’s and iStock’s watermarks. This finding is particularly important because it establishes that AI developers may not be able to escape liability for outputs that reproduce trademarks or distinctive visual elements by claiming that individual users alone are responsible for the generated content.
The judgement represented a setback for photographers and content creators hoping for clear copyright protections. Despite this outcome, Getty and other creative organisations have called on the UK government to strengthen copyright protections and provide greater support for content creators.
The Copyright Law Challenge
The fundamental issue hinges on whether AI training constitutes fair use or copyright infringement. Under current UK law, companies generally need permission from copyright holders to use their work for AI training, except in limited research or study contexts. However, without an established licensing framework and faced with the impracticality of obtaining millions of individual permissions, AI companies have largely proceeded with broad data scraping.
The situation is equally complex internationally. In the United States, technology companies argue that their AI technologies utilise copyrighted works in a manner that qualifies as fair use—analysing materials to produce new and innovative content. Content creators counter that this practice unlawfully replicates their creations to produce competing materials that threaten their income. If courts ultimately rule that such use constitutes fair use, AI companies could potentially escape copyright liability entirely.
Pinsent Masons reports that a German court recently ruled in a different context that a not-for-profit organisation could scrape copyright-protected images for AI training under an exception provided in German and EU law for scientific research. This decision reveals the patchwork nature of international copyright law and how different jurisdictions may reach conflicting interpretations.
The Impact on Photographers
The consequences for photographers are tangible and measurable. When the Association of Photographers (AOP) surveyed its members in September 2024, 30 percent reported losing commissioned work to generative AI. By February 2025, just five months later, this figure had increased dramatically to 51 percent, representing an average value loss of £14,400 per professional photographer and a total membership loss of approximately £43.2 million. (Source: The British Journal of Photography).
This economic impact extends beyond lost commissions. Photographers worry about brand dilution, where AI systems trained on their style can produce outputs that mimic their distinctive artistic approach, flooding the market with cheaper alternatives. Additionally, there’s the concern that AI-generated imagery using their visual language could compete directly with their professional services.
Copyright Protection for AI-Generated Works
An important related question is whether photographers using AI tools can themselves claim copyright protection for the resulting images. The answer is nuanced and jurisdiction-dependent.
In the United States, the Copyright Office has established that works created solely by AI without human involvement are not eligible for copyright protection. However, if a photographer contributes significant creative input—making meaningful choices about composition, editing, or directing the AI’s output—those elements may be protected.
The European Union and UK take slightly different approaches, generally requiring human authorship but potentially allowing copyright protection if the photographer demonstrates substantial control over the creation process. However, the threshold remains high, and courts have been sceptical of claims where AI bears substantial responsibility for the final output.
What Protections Are Available?
For photographers concerned about their work being used for AI training, several strategies exist, though none are foolproof:
Opt-out mechanisms: Some AI companies, including OpenAI, have introduced tools allowing creators to exclude their work from future training datasets. OpenAI provides methods to prevent GPTBot from accessing websites and offers processes for photographers to opt out of DALL-E training. However, these are reactive measures and don’t prevent historical use of already-published work.
Watermarking and metadata: Strategic watermarking and embedding copyright metadata in images can provide some deterrent to unauthorised use, though determined data scrapers may remove or ignore these markers. The UK government consultation has suggested that watermarking might become part of a broader protective framework, though this would shift the burden of protection onto photographers rather than AI companies.
Legal registration: Ensuring photographs are properly registered with copyright offices can strengthen legal claims and enable statutory damages if infringement is proven. In the US, registered works can result in statutory damages ranging from $750 to $150,000 per work for wilful infringement.
Platform choice: Choosing storage and hosting solutions that explicitly guarantee data won’t be used for AI training can provide peace of mind, though these are often premium services.
The Regulatory Landscape and Future Direction
The legal and regulatory environment is rapidly evolving. The UK government launched a consultation in December 2024 on copyright and AI that closed in February 2025, proposing four options for copyright protection. These ranged from maintaining the status quo to strengthening copyright through mandatory licensing—an approach that would require AI developers to obtain and pay for licenses to use copyrighted material.
A significant development came in January 2025 when the House of Lords voted to support amendments to the Data (Use and Access) Bill specifically designed to bolster creator protections. This legislative momentum suggests that copyright law may be moving toward requiring explicit licensing and compensation for AI training use.
Moving Forward: What Photographers Should Know
The conflict between AI images and photographer copyrights remains unresolved in many jurisdictions. While major legal cases continue to unfold, photographers face a dual challenge: protecting existing work from unauthorised AI training while understanding their own rights when using AI tools in their practice.
Key takeaways
The primary concern is that vast amounts of copyrighted photography have already been used to train AI systems without permission, creating a significant legal and economic disparity.
Copyright protections remain unclear but are generally stronger under UK law than in some other jurisdictions, though recent court rulings have been mixed.
Individual photographers have limited practical ability to prevent web scraping without government-level enforcement of stronger copyright protections.
Professional photographers should document their work, maintain proper registrations, and stay informed about platform terms of service.
The industry is mobilising for stronger protections, with organisations representing over 100,000 visual creators calling for retrospective settlements, transparent disclosure of training datasets, and fair licensing agreements.
As this legal landscape continues to evolve, collaboration between photographers, regulators, and technology companies will be essential to ensuring that the benefits of AI innovation don’t come at the expense of photographers’ intellectual property rights and livelihoods.
For West London Studio’s perspective: As creative professionals, the issues affecting individual photographers directly impact studios like ours. Understanding these copyright challenges helps us advocate for fair practices, ensure our licensing agreements reflect the evolving AI landscape, and help our clients navigate the complex questions surrounding AI-assisted creative work. The future of photography depends on establishing clear, fair rules that respect creator rights while allowing beneficial innovation to flourish.