The Government’s plans for AI and copyright have been further thrown into doubt as two of the leading developers of artificial intelligence – OpenAI and Google – have poured cold water on the Government’s preferred approach in their respective responses to the consultation on AI and copyright (which we have discussed here).

We have written previously (see, for example, here) about the challenges facing the Government in trying to square the circle of protecting the rights of copyright holders whilst also allowing AI developers to train their models on massive datasets which often include copyright-protected material. In the consultation, the Government proposed its preferred approach which would see the introduction of an opt-out system – similar to that in the EU DSM Copyright Directive – whereby the text and data mining (“TDM”) exception under the Copyright, Designs and Patents Act 1988 would be expanded to permit data mining on copyright-protected works for commercial purposes, but allow right holders to opt out by expressly reserving their rights.

Responding to the consultation, OpenAI has said that it does not agree with this approach. Instead, it believes that a broad text and data mining exception, without an opt-out system, is preferable. It explains its position by pointing to the “challenges of implementing opt out systems” as exemplified by experiences in the EU where “the lack of clear and scalable technical standards has created uncertainty about what opt-out methods are workable and valid, causing uncertainty for both AI companies and rightsholders”. More fundamentally, OpenAI stresses that a policy framework must recognise “the crucial role of data in developing advanced AI systems” and that the UK must create a “clear, predictable regulatory environment that sets it apart from other jurisdictions to boost its competitiveness”.

Whilst Google’s response agrees that the position that OpenAI has adopted “is the most competitive option and would support the Government’s ambitions to be an AI leader”, it does not reject outright a mechanism for rightsholders to reserve their rights. However, it states that the use its models make of data is “non-expressive” and, as such, should not be covered by the exclusive rights of the rights holder. Similarly, it “reject[s] the framing of rights reservation as a licensing mechanism”, believing that training on the open web must be free. It also expresses concerns that “the excessive transparency requirements [in the Government’s preferred option] could hinder AI development and impact the UK’s competitiveness in this space”.

At the same time as the publication of the submissions from OpenAI and Google, the Tony Blair Institute for Global Change published a paper, “Rebooting copyright: How the UK can be a global leader in the Arts and AI”, which supported a broad TDM exception whilst pointing out that an opt-out policy could be “challenging”. Similarly, the Chair of the Startup Coalition has written to the Culture, Media and Sport Committee expressing the frustrations of a number of AI startup founders at the regulatory environment in the UK.

These views stand in marked contrast to the outcry from those in the creative industries who have urged the Government to rethink its plans. We have previously commented (see here) on, among others, members of the film, television, publishing, theatre, and music industries calling on the Government to change its plans. Recently, a number of high profile architects, designers, and members of the fashion industry also added their voices to the chorus.

As for where we go from here, the Government has stated that it has received 11,000 responses to its consultation and is “taking care to read each response”. Whilst some recent reports have suggested that ‘concessions’ have been offered to the creative industries (such as preparing an economic impact assessment on its proposals), the official line is that no decisions have been taken, and that we can expect a response to the consultation later in the year.

To read the submissions from OpenAI and Google, click here and here.