shutterstock_638342005_tatiana_shepeleva
19 June 2019Copyright

Is it time for IP monopoly to come to an end in AI datasets space?

Many more questions were raised than answers provided yesterday at the “AI: decoding IP” conference, perhaps a reflection of the speed at which artificial intelligence technology is outpacing legislation and disrupting industries.

Lord Tim Clement-Jones, chair of the House of Lords AI select committee, said: “My colleagues in [UK] parliament are beginning to get a glimmer of the ethical issues and some regulatory issues, such as facial recognition which may require regulation in the short term, [in the AI space] but as far as IP issues go, we’re going to have to lead them extremely gently by the hand for the foreseeable future.”

Clement-Jones was speaking yesterday at the “AI: decoding IP conference”, hosted in London by the UK Intellectual Property Office (IPO), in cooperation with the World Intellectual Property Organization (WIPO).

During the panel “Applications of AI and new business models”, Clement-Jones asked: “How do we reconcile the desire to protect the IP in datasets while resisting monopoly power and loss of public benefit? Does and should copyright unduly limit the availability of datasets for use of for things such as training?”

Deep neural networks rely on being fed very large data sets for training purposes, but there is a risk of copyright infringement.

While there are some defences, such as where this a temporary copy of the work or fair dealing with the work for the purposes of research for non-commercial purposes, Clement-Jones warned that a commercial purpose or value will often be contemplated, even if indirectly, by those using the datasets to train their networks.

This, he said, raises the question of whether the EU Directive 1996/9/EC (the Database Directive) and the UK legislation implementing the directive (Copyright and Rights in Databases Regulations 1997) are fit for purpose anymore and, even if they are, whether current text and data mining exceptions are sufficiently broad.

Under current EU law (as the EU Directive on copyright in the Digital Single Market has not yet been implemented in member states), researchers intending to mine works protected by copyright would need authorisation from the database owner or an appropriate copyright exception, and there is much uncertainty.

Clement-Jones added: “Should [the exemptions] cover commercial applications too? Is training really copying in these circumstances? Alternatively, should we be regulating to ensure access to some data sets under compulsory creative commons licences [a copyright licence that allows for the free distribution of the otherwise copyright-protected work], … but perhaps with the addition of a permission to use for certain commercial uses or public benefit.”

Concluding his presentation, Clement-Jones elaborated on the question of distribution of benefit, a question he said was very much linked to the question of reward for IP creation.

“If AI creates huge quantities of IP autonomously, who gets the benefit? Big corporations that have access to vast quantities of data to train algorithms? Or should the age of the IP monopoly come to an end for certain purposes? Should our competition regulators intrude on IP monopolies in the way they have never done before?” he concluded.

Sitting on the later panel “AI & IP – Disrupting the established”, Belinda Gascoyne, senior IP law counsel at IBM Europe, Middle East & Africa, added that, in terms of copyright law, there is a need for laws in all jurisdictions to be urgently reviewed and updated.

“The technology is moving incredibly fast, while creating legislation is slow. It’s difficult for the legislators to keep track,” she said.

Gascoyne welcomed the mandatory text and data mining exceptions in the digital environment in the controversial EU Directive on copyright in the Digital Single Market. The new directive has confirmed that research organisations and cultural heritage institutions are exempt for the sole purpose of scientific research, providing legal certainty to these entities.

The second exemption expands the exception to any entity, but permits copyright owners to opt-out of the application of the exception.

However, Gascoyne warned that the directive is being challenged by the Polish government and member states are yet to transpose it. She added: “Giving practical advice on what clients can and can’t do [in this space] is quite impossible.”

Rewarding innovation

Turning to patents in the second session, “ AI & IP – Disrupting the established”, Ryan Abbott, professor of law and health sciences at the University of Surrey School of Law in the UK, said that we “shouldn’t just try and shove the AI phenomena into existing laws”.

He added that we need to think about why we have patent protection—historically it’s been an innovation incentive.

“If what we want from patent law is innovation, investing in inventive machines is going to give us that,” said Abbott.

According to the professor, the fairest and simplest way to do this is to acknowledge machines as inventors where a person doesn’t meet the inventorship criteria (for example, IBM would own the patent and IBM Watson, a supercomputer that combines AI and analytical software, would be listed as the inventor).

Although change “seems to be coming fast” in the patent field, said Gascoyne, citing the European Patent Office’s recently issued extended guidelines on AI and proposals to reform patented subject matter in the US, there is a lack of clarity on what exactly reform will entail.

“Five years ago in Europe, the patentability of computer programmes seemed to be set in stone.  Suddenly, within the last two months we’ve seen a referral to the European Patent Office’s (EPO) Enlarged Board of Appeal,” added Gascoyne.

In March this year, the EPO’s Technical Board of Appeal referred questions relating to computer-implemented inventions to the Enlarged Board of Appeal (EBA). The case concerns a simulation run on a computer, and the patentability of such simulation methods.

Abbott concluded: “First of all, businesses like IBM need certainty. Countries need to spend time properly discussing these [IP] issues and there needs to be a broad base consultation. Ideally legislative bodies will pass new laws to ensure we’re going to get the most benefits out of AI that will benefit the most number of people.”

Did you enjoy reading this story?  Sign up to our free daily newsletters and get stories like this sent straight to your inbox.

Today's top stories:

EU General Court rules Adidas three-stripe TM is invalid

AC Milan mark survives Marriott opposition

US bill seeks to block Huawei from seeking patent relief

UK Justice hails profound promise of AI

Already registered?

Login to your account

To request a FREE 2-week trial subscription, please signup.
NOTE - this can take up to 48hrs to be approved.

Two Weeks Free Trial

For multi-user price options, or to check if your company has an existing subscription that we can add you to for FREE, please email Adrian Tapping at atapping@newtonmedia.co.uk


More on this story

Copyright
18 June 2019   The late Steven Hawking described artificial intelligence as “the best or worst thing to happen to humanity”, while Tesla founder Elon Musk claimed that with the use of AI “we are summoning the demon”.