The U.K. is planning to tweak an existing law to allow text and data mining “for any purpose,” in a move that’s designed to boost artificial intelligence (AI) development across the country.
The announcement constitutes part of a broader strategy to “level up” AI and transform the U.K. into what it calls a “global AI superpower” — and part of this will involve reassessing existing intellectual property (IP) laws. Following a two-month consultation period where stakeholders from across the industrial spectrum were asked for input, including rightsholders, academics, lawyers, trade organisations and businesses, the U.K.’s Intellectual Property Office (IPO) today published its response and confirmed what will (and won’t) be changing moving forward.
Algorithms
Text and data mining (TDM) is pivotal to the development of new AI applications, allowing researchers and businesses to copy and harness disparate datasets to train their algorithms. However, gaining access to enough relevant data has inherent challenges — the data is often owned by third-parties that may only want to make data available under a commercial license, if they make it available at all.
Back in 2014, the U.K. amended its existing TDM regulations — which are related to the broader Copyright, Designs and Patents Act (1988) — to include an “exception” allowing AI researchers to leverage third-party data for non-commercial purposes without incurring major costs. However, this still imposed significant limitations on how the data could be used and discouraged businesses from investing in AI development. Moreover, it didn’t extend to database rights, which is distinct from works covered by traditional copyright law.
Today’s announcement essentially resolves this. The U.K. government is now planning to adopt a TDM exception that covers any purpose far beyond research and academia, with no opt-out for rightsholders. On top of that, it also includes provisions for database rights.
This runs contrary to the comparable European Union (EU) Directive on Copyright in the Digital Single Market, which offers a mandatory exception only for TDM in the scientific research domain. Indeed, rightsholders can opt-out their copyrighted work from commercial use cases, meaning they can still monetize TDM.
The U.K.’s proposed changes could serve as a significant component of its stated “levelling up” plans, given that access to AI training data is a major stumbling block for all but the largest companies. But more than that, it seems that the intention here is to lure AI companies to the U.K., knowing that they have more freedom to conduct text and data mining. This is particularly important in a U.K. that is now in direct competition with the EU following its exit from the bloc in 2020, as it acknowledges in today’s published response:
These changes make the most of the greater flexibilities following Brexit. They will help make the U.K. more competitive as a location for firms doing data mining.
The Brexit effect
This new copyright and database rights exception, which the government plans to enshrine in “suitable legislation” in due course, effectively moves the balance of power away from rightsholders and heavily toward businesses and other commercial entities. But the transition could have unintended consequences, according to some. Under the proposed new rules, the data-miner still has to acquire the data through lawful means, meaning that it has to be publicly available (e.g. as part of a subscription). So rightsholders who may previously have charged for TDM as part of a data-licensing model may instead withhold their data altogether, which could adversely impact AI development in the future.
“On the one hand this decision can be seen as an enabler for AI development, but on the other it may have the effect of encouraging copyright holders to place more restrictions on their content,” Richard Johnson, partner at European IP law firm Mewburn Ellis, told TechCrunch in a statement.
It’s worth noting that rightsholders will still have some rights themselves under the new rules, in terms of where they choose to publish their data or copyrighted works, and they can still charge for access to that data. They just “will no longer be able to charge for U.K. licences for TDM” specifically, and they won’t be able to enforce any kind of opt-out — any entity that legally accesses the data will be able to mine it.
Status quo
The U.K. government’s published response today is notable not only for what’s changing with text and data mining, but also for what isn’t changing. Central to the consultation was whether computer-generated works (CGWs) without a human author should continue to be protected under copyright law — the U.K. is actually one of the few countries that bestow copyright protection on CGWs, with a 50-year protection period in place (this compares to 70 years for human-generated works).
While the U.K. sought input on whether the protection period should be altered or scrapped altogether, it ultimately decided not to change anything, noting that the existing protections around CGWs were not “harmful” and that the use of AI is still in its relative infancy. “We will keep the law under review and could amend, replace or remove protection in future if the evidence supports it”, the response noted.
Similarly, the U.K. has also decided that AI systems still can’t be given patents for inventions, despite some early movements in that direction in a couple of jurisdictions around the world. The main reasons it cited were that AI isn’t advanced enough yet to “invent” without significant human input — a legal opinion that is shared by most nations. In its consultation response, the U.K. said that it was reluctant to diverge too much from “international norms” on inventorship.
“This may be seen by some as a lost opportunity — however, the commitment to active engagement in working towards international consensus is very positive to see,” Johnson said. “From an end-user perspective, it would be undesirable for the IP landscape to be a fragmented on this issue. The government response is therefore in line with many of the court decisions in this area, in that it recognizes the potential for change in future, but avoids any immediate action.”