The U.K. plans to amend existing law to allow text and data mining “for all purposes.” This move is intended to encourage the development of artificial intelligence (A.I.). All across the country.
This announcement is part of a larger strategy to “level up” A.I. The U.K. will become a “global A.I. superpower. After a two-month consultation period in which stakeholders from all sectors were asked for their input, including rights holders and lawyers, as well as businesses and academics, the UK.
Intellectual Property Office (IPO) today published its response. It also confirmed what would (and wouldn’t) change moving forward.
AI Development by eliminating Data Mining
Data mining (TDM) and text and data mining (TDM) is crucial to developing new A.I. Applications that allow researchers and businesses to access disparate data to train their algorithms.
Accessing enough data can present challenges, and data is often held by third parties who may not want it made available without a commercial license.
The U.K. changed its TDM regulations in 2014, which were related to the broader Copyright, Designs, and Patents Act (1988), to add an “exception” that allows A.I. Researchers can use third-party data without high costs for non-commercial purposes.
This imposed restrictions on the use of data and discouraged businesses from investing in A.I. development. It did not extend to database rights which are different from works covered under traditional copyright law.
Today’s announcement essentially resolves this. The U.K. government will adopt a TDM exemption for all purposes, including research and academia. Rightsholders can opt out, and it also contains provisions regarding database rights.
This is contrary to the E.U. (European Union) comparable. This contradicts the European Union (E.U.) directive on Copyright in the Digital Single Market.
The only exception to the directive on Copyright in the Digital Single Market is TDM in scientific research. Rightsholders can opt out of copyrighted work for commercial use, so they can still monetize TDM.
#2. Brexit: The effect
Given that the U.K. has access to A.I., these proposed changes could form a major component of its “leveling up” plans. Training data is a significant obstacle for any company, even the largest.
More than that, this is a way to attract A.I. Companies will flock to the U.K. because they know they have greater freedom to do data and text mining.
This is especially important for the U.K., which is currently in direct competition with the E.U. As acknowledged in today’s reply, the U.K. will leave the bloc in 2020. The government intends to make this new exception for copyrights and database rights in “suitable legislation” over time.
This will effectively shift the power balance away from rightsholders to businesses and other commercial entities. Some believe that the transition could have unintended effects.
The new rules propose that the data-miner must still acquire the data by lawful means, which means it must be made publicly available (e.g., as part of a subscription). Rightsholders previously charged for TDM in a data-licensing arrangement may now withhold their data, which could negatively impact A.I.’s Future development.
“On one hand, this decision could be seen as an enabler of A.I. Richard Johnson, partner at European I.P., said that while this decision can be seen as an enabler for A.I., on the other hand, it could have the effect of encouraging copyright owners to place more significant restrictions on their content.
Mewburn Ellis, a law firm, informed TechCrunch. Notably, rightsholders will retain some rights to choose where to publish their data and copyrighted works. They can still charge access to the data.
They won’t have the ability to set for U.K. TDM licenses, and they won’t be able to enforce any opt-out any entity legally capable of accessing the data will be allowed to mine it.
#3. Status quo
Today’s U.K. government response is noteworthy for its published responses. It outlines what’s changed with text and data mining and what isn’t.
The consultation focused on whether computer-generated works (CGWs), without the involvement of a human author, should be continued to be protected by copyright law. The U.K. is one of the few countries that grants CGWs copyright protection within 50 years (compared to 70 for human-generated works).
The U.K. requested input about whether to alter or eliminate the protection period. However, it decided not to change anything. It noted that existing CGW protections weren’t “harmful” and that A.I. was unnecessary.
It is still very young. The response stated that the law would be reviewed and could be amended, replaced, or removed in the future if there is sufficient evidence. The U.K. also stated that A.I. was not allowed to be patentable.
Despite some initial movements in this direction in a few jurisdictions around the globe, patents cannot be granted to invention systems. It cited A.I. as the main reason. A.I. is not advanced enough to “invent” with minimal human input, a legal opinion shared across most countries.
The U.K. responded to the consultation by saying it would not deviate too much from “international norms” on inventorship. Johnson stated that while some may view this as a loss of opportunity, Johnson believes the commitment to active engagement in working towards international consensus was very positive.
It would be unwise for the I.P. to do so from an end-user point of view. This issue is not well-defined. In that the government acknowledges the possibility of change in the future but does not take any immediate action, it is in line with other court decisions in the area.