»ÆÉ«¶ÌÊÓƵ

AI and intellectual property

Abstract: This seminar will review selected key dilemmas in AI and intellectual property protection, covering the current debate on copyright 'text and data mining' (TDM) exception and the disclosure requirement in patents protection. 

The first part of the presentation will focus on copyright of training data. The development of generative AI has generated countless discussions about its impact on society and the economy. One of the disputed aspects is the use of various materials, such as texts, images, sounds and videos , for training these systems. Copyright legislation in the UK (and the EU) contains specific exceptions for the use of copyrighted works in 'text and data mining' (TDM) activities that might be used in this context under some limited circumstances. The government has recently proposed to introduce further flexibility into these norms to facilitate AI development, which has been met with strong opposition by the creative industries. We will discuss the TDM exception, how it came about and its use and challenges, including the current debates on expanding it.

The second half of the presentation will review the evolving landscape of key legal elements in industrial property protection, including trade secrets protection and the patentability criteria (novelty and inventive step) due to the explainability/transparency requirement in AI regulation.  

Talk delivered on 23rd April 2025 as part of »ÆÉ«¶ÌÊÓƵ AI seminar series. Watch the recording on