Online only

Unlawful use of creators’ content by Big Tech must end

[Creators' Rights Alliance]

THE CREATORS' Rights Alliance (CRA) has written to a range of software companies reaffirming that members of the Creators’ Rights Alliance and the 500,000 creators they represent do not authorise the use of any of their works for the training, development, or operation of AI models. The National Union of Journalists is a founder member of the Alliance. The companies include Microsoft, Google, OpenAI, Apple and Meta - owner of Facebook, Instagram and WhatsApp.

In the letter the CRA "notes with concern the use of vast amounts of work protected by copyright and related rights without the authorisation of the rightsholder(s) and creator(s) in the development (including so-called training), and operation of AI models. This has caused and continues to cause great harm to the significant creative, human, and financial investment made by authors, performers, and visual creators."

In announcing the letter the CRA "encourages all creators to make use of the Intellectual Property Enterprise Court to combat infringement, particularly the small claims track."

Among the demands that the CRA makes of the tech companies are "to provide full transparency about the works which have been used to develop their model"; "to engage in good faith licensing negotiations to redress past bad practice, to remove from their systems any copyright protected work... which has been used without authorisation"; and "to safeguard human creativity, truthful content and the rights of authors, creators, and performers, it is vital that AI models are developed and used in a legal, sustainable, and ethical manner."

Elsewhere in the ‘AI’ woods...

The Society of Authors (SoA) reports concerns that "a number of publishers, including Taylor & Francis (Routledge), Wiley, Oxford University Press and Cambridge University Press have licensed rights or are contemplating licensing rights in some or all of their catalogue of titles, to AI companies." The SoA said of an earlier report about Taylor & Francis that such uses will "harm the quality, integrity and essential variety of materials being used in education and academia, as these would be created without the safeguards and context of clear provenance".

The initial fee that Taylor & Francis is reported to have agreed is $10 million. The Freelance sees that as extraordinarily low and suspects that these academic publishers' practice of demanding assignment of all rights in what they publish essentially for free contributes to them under-valuing it.

The SoA, representing book authors, suggests that "If you believe your work has been included in an AI licence without your consent or if you have been contacted by your publisher with any changes to existing terms and conditions, please get in touch." Please also inform the Freelance.

Even more urgently, let us know about any deals that newspaper and magazine publishers may be striking over your head. On 4 June representatives of the Guardian and the Sun told a business conference that they were willing to do deals with AI companies "on the right terms". The Sun typically demands of journalists a licence to do everything with their work. It is not clear to the Freelance whether the standard Guardian licence allows it to collect cash for ingestion of freelances' work into AI systems.

Guardian Media Group chief executive Anna Bateson did say: "I’m sure we’d all love to say: ‘We’re going to reduce our cost to our businesses by not paying for the things that are fundamental to the way that we run our businesses.’ We’d love not to have to pay for web hosting, we’d love not to have to pay for print or paper... a respect for copyright" would have to "underpin any deal".

Then on 7 August the Press Gazette reported that the Financial Times, Fortune, German tabloid publisher Axel Springer and The Atlantic have all agreed to license their content to Prorata.ai. This startup declares that its "mission is to ensure that publishers and creators are fairly compensated by generative AIs".

Monopoly misery

The difficulties that new entrants such as Prorata.ai will face in elbowing into the market are highlighted by a court judgment on 6 August. The US District Court for the District of Columbia ruled that Google has broken competition law "by maintaining its monopoly in two product markets in the United States: general search services and general text advertising". As Brian Fung and Clare Duffy report for CNN, when giving evidence in the case Microsoft chief executive Satya Nadella observed that Google's monopoly of search services threatens to give it an unbeatable advantage in AI that would further entrench its power.

The court has not (yet) imposed a penalty. Judge Amit Mehta did offer an extensive discussion of the past failure of simply showing users an alternative to a monopoly product. Back in 2001 the legal requirement for users to see a "choice screen" before installing Microsoft's web browser program Internet Explorer failed to have much impact, as did similar measures in 2018 in the European Union. It does seem likely to the Freelance that the conclusion of this process must be the breaking-up of Google.

We now remind ourselves, however, that in an earlier stage of US v Microsoft a lower court in 2000 ordered a breakup of Microsoft; and that this was reversed on appeal. In any case, if one of the components of a Google breakup retained a dominant position in the search business, it would retain the same advantage in AI training that Brian Fung and Clare Duffy identify.


9 September 2024 For the record: on 20 August Condé Nast, publisher of Vogue and Wired, announced a "multi-year deal" with OpenAI. No word yet on how little cash they're getting. Wired reported the comment: "no one wants to help train the tools spreading misinformation and degrading the skills many of us spent decades honing,” from one writer for a Condé outlet, who requested anonymity out of concerns for professional reprisals.