Copyright conflicts coalesce

GOOGLE and Microsoft have issued reports, both entitled Unlocking the UK’s AI potential, that set out their stall for government action to boost their profits - sorry, to "harness AI for economic growth" as the version commissioned by Microsoft puts it. The Google version calls on government to "enable TDM for both commercial and research purposes": "TDM" is "text and data mining".

Image © Microsoft

Both hold out the carrot of investment in data centres and thus of some jobs. The Microsoft report focuses on possible savings in the public sector - and on getting access to data held by government (and by implication by the National Health Service).

That "enabling" requested by Google would mean legislation to introduce a new "exception" to UK copyright law - a rule allowing copying without permission of the author (or other holder of copyright).

Clearly, there is a risk that ministers will be dazzled by the promise the tech giants hold out; and equally clearly we will resist this, working with allies for example in the British Copyright Council.

In doing so we will learn from what went wrong in the European Union, particularly that "safeguards" on such measures can easily be ineffective. In 2019 the EU passed legislation permitting TDM for not-for-profit research. The owners of the LAION dataset of images promptly claimed to have gathered these for non-profit research - and supplied them to train machine-learning image generators that are moving into for-profit activity: see here.

Monopoly challenge

Meanwhile in the Southern District of New York, six academic publishers have been hit with a lawsuit claiming that they broke US competition law - known there as "anti-trust" law.

The claim is that the publishers colluded to keep the rate of pay for peer-review of articles at zero; that by requiring researchers to submit articles to only one journal at a time they ward off competition between their journals; and that their policies prevent researchers sharing their work while it undergoes review. The Freelance is puzzled that it does not headline the quite remarkable similarity between the publishers' policies requiring researchers to assign all rights in their articles and papers to the publishers - but it is dealt with.

The case is brought in the name of Lucina Uddin, a brain researcher at the University of California, Los Angeles (UCLA). The aim is to get court permission to convert it into a "class action" - a process in US law that in this instance would allow all academics to join the "class" and, if successful, share in compensation.

© Mike Holderness

'Artificial Intelligence' our coverage to date