What Designers Need to Know About the Latest Legal Changes
An increasing number of designers are working with artificial intelligence (AI). Images can be created in seconds, product ideas can be visualised with just a few prompts, and initial designs can be developed more quickly than ever before. However, as these new tools become more seamlessly integrated into the design process, legal questions arise:
- Who owns a generated design?
- When must the use of AI be disclosed?
- And what responsibility do companies bear when technology becomes a creative tool?
These were the questions addressed at the Members Talk event organised by the German Design Council Foundation. During this online discussion, Dr Christoph Matras and Sarah Rost, lawyers at the commercial law firm FPS, highlighted the biggest legal pitfalls and explained why companies should address compliance issues before the finished product is ready.
Which AI tool you choose is more than just a practical decision. When choosing between ChatGPT, Claude, Midjourney and other applications, it is important to consider not only which system delivers the best results, but also the terms of use and whether the generated output can be used commercially. Added to this are the requirements of the EU AI Act, which is gradually introducing transparency and documentation obligations. “We need to factor in the commercialisation of AI-generated designs right from the start of the creative process, not just when the final product is ready,” explained Sarah Rost. The legal review begins much earlier than many companies realise.
It’s the Human Touch That Makes the Difference
Copyrights for AI-generated works remain a key issue. ‘Works generated purely by AI are not eligible for protection. Ultimately, it comes down to the original creative contribution,’ explains Christoph Matras. Therefore, the decisive factor is the creative contribution made by a human being. Under certain circumstances, anyone who selects variants, alters proportions, or creatively refines a design may still be able to establish protection.
There is also frequent uncertainty regarding future labelling requirements. A product designed with the aid of AI does not, as a matter of principle, have to be labelled ‘AI-designed’. However, in the case of AI-generated synthetic media that appear realistic, such as images or videos, as well as content intended to influence opinion, Article 50 of the AI Act — the transparency requirement — applies. From 2 August 2026, such content must be clearly labelled as AI-generated and this labelling must be immediately visible.
Article 53 addresses a different issue; it concerns the providers of generative AI models rather than the labelling of individual pieces of content. Among other things, they must disclose information on training data and copyright. This information is particularly relevant for businesses when selecting an AI tool and conducting risk analysis.
Legal Certainty Requires Clear Processes
As generative AI becomes more prevalent, the importance of having proper contracts and internal processes in place is also increasing. Companies should document the tools they have used, record human creative input in a traceable manner and set out responsibilities in contracts. Ultimately, it is not the AI that remains liable, but the individual or company that publishes the output.
Another aspect that will become increasingly important in future is AI competence. The AI Act stipulates a duty to ensure sufficient AI expertise. Companies must ensure that their staff using AI systems are knowledgeably competent. This involves more than just confidently using prompts; it also requires an understanding of the legal, technical, and creative limitations of the systems.
The key message of the Members' Talk was therefore that, while generative AI opens up new possibilities, it does not replace creative decision-making or legal responsibility. Those wishing to harness the technology's potential should therefore not view it in isolation, but as part of a professionally organised design process — from tool selection to marketing.
What Is the EU AI Act?
The EU AI Act is the first comprehensive piece of legislation to govern artificial intelligence in Europe. It classifies AI applications into three risk levels: systems posing an unacceptable risk, such as state-run social scoring schemes, are prohibited. High-risk applications, such as programmes for the automated assessment of job applications, are subject to strict requirements. All other AI applications remain largely unregulated. While generative AI is not considered a high-risk system, it is subject to labelling and transparency requirements.
Further links:
FPS’s ‘AI Check’ provides a concise overview of the key provisions and requirements of the AI Regulation: fps-law.de/en/FPS-AI-Check
Official overview page of the European Parliament: www.europarl.europa.eu/topics/en/article/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence
Official website of the European Commission. The Code of Practice helps the industry to comply with the legal obligations of the AI Act regarding safety, transparency and copyright for general-purpose AI models: digital-strategy.ec.europa.eu/en/policies/contents-code-gpai
