If an investor was sitting across the table from you, asking: “How much of your core product was built using AI?” and “Can you confirm you own it?”, could you answer that with confidence?
For many founders, the honest answer is: not entirely. And that gap between assuming you own what your AI tools create and being able to prove it is one of the most significant emerging risks in founder due diligence. It is the question that could define your valuation.
The ownership problem
The assumption that paying for an AI tool means you own everything it creates is not always correct. Ownership of AI-generated outputs depends on the nature of the output, the contract governing the AI service, the degree of human involvement and the applicable jurisdiction. Different providers approach this differently. Some grant users broad rights over outputs. Others impose restrictions, claim licences or qualify ownership in ways that matter enormously if your business is built on AI-generated content or code.
So what are the IP risks that founders need to understand?
- Protection risk: Works generated by AI without sufficient human creative input may not attract full copyright protection under UK law. If your product relies heavily on AI-generated content, your IP position may be weaker than you think and weaker than you’re representing to investors.
- Infringement risk: AI models are trained on vast datasets. There is a real and growing risk that AI-generated outputs may reproduce or closely resemble third-party copyrighted material, exposing your business to infringement claims you didn’t see coming.
- The training data question: A critical distinction exists between data that is publicly accessible, and data that is free to use for AI training. These are not the same things. Conflating them is a common source of legal risk for businesses building AI products.
WHAT CAN YOU DO?
- Audit the IP clauses in your key AI supplier agreements and understand what rights you actually have over outputs.
- Document meaningful human contribution to AI-assisted work, particularly for strategically valuable assets. Good records become extremely valuable during due diligence.
- Update your IP policies to specifically address AI-assisted creation. Most existing policies predate generative AI.
- Assume investors will ask detailed questions about AI and IP. Develop your answers before those questions arise.
While it may have been the case five years ago that investors rarely asked detailed questions about AI-generated IP, this has now changed. Founders who can explain clearly how key assets were created, what rights exist and how governance works are significantly better positioned than those who cannot.
FURTHER INFORMATION
Our Corporate & Commercial team advise firms on a wide range of AI-related matters. If you would like to know more, please do get in touch with Chris who will be happy to help.
