Two reports on artificial intelligence (AI) published recently by UK Select Committees have highlighted various ongoing challenges for AI policymakers on issues around intellectual property, specifically copyright.
The Science, Innovation and Technology Select Committee recently published its interim report: The Governance of Artificial Intelligence, following its inquiry into the impact of AI on various sectors.
The committee’s high-level report identifies 12 ‘challenges’ of AI governance that policymakers should address to secure public safety and confidence in AI. One key challenge is the ‘intellectual property and copyright challenge’, which highlights the importance of protecting the rights of content creators whose copyrighted works may be used by AI models and tools.
The report flags that concerns have been raised by some in relation to the practice of web ‘scraping’. This is where copyrighted works obtained from public online sources are fed into an AI model that ‘learns’ from factual and statistical features of those works. This learning is used to generate new works.
The report indicates that some in the creative industries have expressed a desire to collaborate with the AI sector to develop a licensing framework for the use of copyrighted materials in training AI. The report also notes that the UK Intellectual Property Office (UKIPO), in discussions with field leaders in various sectors, is looking at developing a voluntary code of practice on copyright and AI. There has not yet been an official update on the progress of these discussions, but for a glimpse at what sorts of issues led to initial calls for a code of practice to be established, see our earlier article on the rise of AI-generated music here.