Distillation: When copyright meets trade secret in an AI world.
If you are attending LESI in Dublin, we look forward to seeing you there.
Last year, in Singapore we co-launched a book on Business Secrets Management, and co-hosted a roundtable: “Business Secrets Management – New Practice Challenges in an AI-Centric World”.
Following the success of that session we are pleased to be co-hosting a roundtable at LESI2026, Tuesday 28th April “Distillation: When copyright meets trade secret in an AI world.” (https://lesi2026.org/program/)
LESI Dublin – What better place to discuss distillation! Although this session will discuss distillation of AI models and other matters relating to the cross-section of AI, copyright and Trade Secrets.
As the legal sector struggles to keep pace with the rapid developments in AI new challenges arise that challenge the ability to protect value creation. In January 2025 the launch of DeepSeek sent shock-waves through the industry first grabbing attention on the size and development cost of training; but then from accusations of obtaining their technology from OpenAI in an unauthorised manner through a technique known as “distillation”.
Whether distillation is a legitimate technical process or mass data misappropriation is a key discussion point. The AI industry actively argues its position on copyright, claiming that training AI models on copyrighted material without a license is fair use.
In light of this the session will particularly focus on the issues around the available mechanisms to protect, and consider whether it is hypocrisy to invoke fair use to defend your own behaviour, while invoking trade secret and contract law to protect your own models against replication or distillation; or is it sound strategy?
- Whether the outputs of AI models qualify as creative expression and hence qualify for copyright protection
- Where the discussion on fair use of copyright hits meets trade secret violation
- The possibility of non-IP approaches such as violation of contracts, user agreements, or Terms of Service
- Contractual arrangements and the principle of good faith in business interactions
