Your rights are under threat from AI. It’s time to have your say!

Last year, 13,500 ALCS members told us what they thought about AI, for which we’re very grateful. The responses we received will shape our contribution to the government consultation and any future licensing plans we pursue. The responses showed that this is an issue that many of you are concerned with, so if you feel strongly about this, there are a couple of additional things you can do to help ensure creators’ voices are heard.

The Government’s preferred solution is to introduce a copyright exception for the use of works to train AI models, with the ability of rightsholders to “reserve their rights” or “opt their works out”. We believe that this option is inadequate, unworkable and unbalanced against the interests of creators. If you feel the same, here’s what you can do to have your say:

1.Fill in the consultation (before the deadline on 25 February)

It’s not just for organisations to fill in, you can have your say too. We’ve provided guidance to help you complete it below. You can complete it online or by emailing a written response.

2.Write to your MP. 

The more MPs know that know their constituents care about an issue, the more likely they are to ask questions and hold the Government to account. We’ve put a template letter for you to amend on Writers Write – our online platform to help you write to your MP.

The Creators Rights in AI Coalition (of which we are a member) have also put together a letter writing campaign here – Safeguard Creative Rights in AI.

Guidance for completing the consultation

The consultation on AI and copyright is also open for individuals to submit their response. The simplest way to do this is by completing the online survey. If you prefer, you can also access the questions on this PDF and submit your answers by email to copyrightconsultation@ipo.gov.uk.

There are 49 questions in total. However, you do not need to answer every question. Just answer the questions that you feel strongly about. We’ve highlighted below what we believe the most important questions are and provided what we think the relevant points to raise are.

You don’t need to complete everything in one go. Your progress can be saved and you can return to your answers later.

Questions 1 – 3 (online survey)

You will need to provide your name, your email address and the organisation you represent. When asked which organisation you represent, you can simply state your occupation, for example “author”.

Question 4 (online survey) / Question 1 (PDF version) – Do you agree that option 3 – a data mining exception which allows right holders to reserve their rights, supported by transparency measures – is most likely to meet the objectives set out above?

We believe the answer is a firm no. Our concerns around this option are:

• A model which requires individual creators to “opt-out” is both deeply unfair and unworkable from a technical perspective.
• This option puts the onus firmly on the creators to reassert their existing legal rights, rather than on the tech companies to ask for permission.
• Creators cannot reasonably be expected to track every use of their work online to enforce their rights by opting-out.
• This is likely to result in many rightsholders having their works unknowingly/unwillingly used for commercial purposes without any renumeration.

Question 5 (online survey) / Question 2 (PDF version) – Which option do you prefer and why?

We believe that ‘Option 1: Strengthen copyright requiring licensing in all cases’ is the best approach to achieve a favourable outcome for creators. However, for this to work effectively, it is essential that this approach is allied with new transparency rules detailing which works are being used for training AI systems. For the following reasons:

• Existing copyright laws do already require AI developers to obtain permission for the commercial use of copyrighted works, the issue is that these laws are not being observed.
• A flourishing licensing market for the use of creators’ works is the best way to ensure that creators rights are respected and that they can secure fair remuneration if that is what they choose, whilst not restricting technological innovation.

Question 6 (online survey) / Question 3 (PDF version) – Do you support the introduction of an exception along the lines outlined in section C of the consultation?

No, we believe that a copyright exception would be unfair to creators such as writers for the following reasons:

• Copyright isn’t broken, it’s just inconvenient for the people that want to use writers’ works for free.
• A copyright exception requires evidence of market failure, of which there is none.
• The Government has not yet provided an opportunity for a functioning marketplace that can deliver growth to develop.
• The tech companies pay for their electricity and their staff, why shouldn’t they pay for the data (writers’ works) that they use too? Broadcasters and publishers are required to pay to access copyright material, why should AI companies be exempt?

Question 8 (online survey) / Question 4 (PDF version) – If not, what other approach do you propose and how would that achieve the intended balance of objectives?

We believe that the following suggestions would constitute the best approach:

• Compensate creators for the widespread, unauthorised, use of their works that has already taken place.
• For the Government to support a framework for a dynamic and adaptable licensing market for the future use of copyrighted works as training data.

This approach gives creators a genuine choice over whether and how their works are used, and provides them with the opportunity to receive remuneration if that is what they choose. At the same time, it provides a means for AI developers to fairly acquire the data they require.

Question 9 (online survey) / Question 5 (PDF version) – What influence, positive or negative, would the introduction of an exception along these lines have on you or your organisation? Please provide quantitative information where possible.

We believe a copyright exception would have a negative impact on authors. Please use this opportunity to outline the ways in which unauthorised uses of your work to train AI may harm or has already harmed your livelihood as a writer.

Question 16 (online survey) / Question 12 (PDF version) – Does current practice relating to the licensing of copyright works for AI training meet the needs of creators and performers?

We believe it is important to note that:

• Rights to use works to train generative AI are a recent and novel concept. Contracts should expressly deal with these rights, setting out the type of uses covered and the remuneration for them. As a novel right, writers should have the choice over how and if they license their works.

Question 18 (online survey) / Question 14 (PDF version) – Should measures be introduced to support licensing good practice?

Yes, we believe the Government should support the development of a framework for an adaptable licensing system.

Our suggestions for this would be:

• New regulations requiring transparency on uses of copyright works for AI training is an essential precursor to a good practice in licensing.
• Support for a centralised online hub providing access for AI companies to a marketplace for permissions, licences and usage rights, akin to a previous Government-sponsored initiative known as the Copyright Hub.
• Upholding existing copyright laws, which forbid the unauthorised use of copyrighted works for commercial purposes, would also do much to encourage the development of a flourishing licensing market.

Question 22 (online version) / Question 17 (PDF version) – Do you agree that AI developers should disclose the sources of their training material?

We strongly believe that transparency to a high standard is essential for any effective solution to this issue. AI developers must be transparent about what works have been used and the nature of such uses.

In our survey of members, you told us that transparency around the use of your works was both severely lacking and desperately needed.

• Greater transparency is necessary to ensure writers know when their work has been used and to subsequently empower them to exercise choice around such uses.

These points are to provide general guidance for submitting a response. If you are a member of a writers’ union, you may wish to consider consulting with them for more specific guidance relating to your sector before submitting your response.

Ed Newton-Rex, the founder of Fairly Trained, a non-profit that certifies generative AI companies that respect creators’ rights, has also drafted template responses to each question in the consultation which you may find helpful when drafting your own responses.