DACS responds to the Government’s white paper on AI regulation on behalf of visual artists
DACS has responded to a public consultation on the Government’s white paper on artificial intelligence (AI) called A Pro-Innovation Approach to AI Regulation released in March 2023.
The white paper set out ambitions for the UK to become an ‘AI superpower’ whilst addressing the risks that AI can pose. It proposes that a regulatory framework for AI will be underpinned by five guiding principles:
- Safety, security and robustness
- Appropriate transparency and explainability
- Fairness
- Accountability and governance
- Contestability and redress
DACS welcomes these principles, but we are concerned that the white paper explicitly excludes the rights of content producers, such as artists, from the scope of the regulatory framework. We consider this needs to be re-examined as artistic works, alongside text and audio, have been used for machine learning and other AI-related purposes often without artists’ consent or a licence.
Head of Policy & International
To ensure a regulatory framework for AI is suitable for visual artists, DACS has set out recommendations to Government that explore and build on the five guiding principles of the white paper:
TRANSPARENCY: AI models, tools and applications must be transparent about what data is being used for machine learning and where this data is sourced. This data includes visual artworks, and it is essential that artists know whether or not their work has been used to train AI. Furthermore, AI-generated outputs should be clearly labelled so. This will benefit public safety and build public trust in AI.
OPT-IN: Artists must be able to meaningfully prevent their works from being used for AI training. Where AI models cannot ‘unlearn’ from data that’s been already used in training, opt-out models will not be sufficient. DACS calls on AI models to be obligated to receive full authorisation and consent before training takes place.
AUTHORISATION: AI models must be built on the basis that those who own images or other data (whether text, audio or personal data) have authorised this use. This means that AI models MUST act within the existing legal frameworks of copyright law and data protection law.
COPYRIGHT LICENSING: There is no need for wide exceptions to copyright law when licensing solutions can be used instead. DACS has almost 40 years’ experience of licensing copyright-protected artistic works for a range of uses and can work together with the AI sector to develop flexible licences.
ENFORCEMENT OF RIGHTS: Licensing models are the best way for artists to enforce their rights. However, where artists do need to use a form of individual redress for copyright infringement, there is a limit to affordable avenues. We call on Government to consider ways to improve access to justice for copyright holders when their work is used in the context of AI, for example within the Intellectual Property Enterprise Court (IPEC).
AN OPTIMAL WORKFORCE FOR ARTISTS: Artists are low-paid creative workers: a recent survey calculated median pay at £12,500 a year and over 70% of artists work as freelancers. Many are concerned that AI could replace them in the workplace, reduce opportunities, or drive down their pay further. We call on Government to take these issues seriously by establishing a freelance commissioner to review working and payment practices for freelance artists, and the impact AI will have on the freelance creative workforce.
ALIGNED THINKING: The UK is not alone in reviewing how AI will transform society. The EU has introduced new AI regulations and other countries are contemplating the same. We encourage Government to learn from countries that have found solutions for ensuring artists are fairly remunerated through efficient mechanisms such as blanket licensing and levies.
FAIRNESS: Human-centred creativity is fundamental to our society. Artists and other creators produce work that challenges our thinking, educates and entertains us, and improves wellbeing. AI-generated works are derivative and based on pre-ingested human creation. We call on Government to take a holistic approach to ensuring that artists and other creators are properly and fairly treated when considering how to advance the AI sector in the UK.
In addition, DACS will publish its own set principles on AI for government to better support UK visual artists and ensure that they can survive and thrive as new technology advances.
