IP and media law: predictions for 2019

Privacy

In the coming year, we are likely to see the concept of what can be protected as "private information" continue to widen. As a result of the Cliff Richard case, which demonstrated that a criminal investigation together with a Police raid can be private, lawyers will feel emboldened to try to protect clients involved or caught up in investigations of all kinds. For example, if law enforcement agencies start using unexplained wealth orders, would an individual served with one be able to prevent their identity being disclosed in the media prior to compliance or judicial challenge?

We can also expect to see individuals try to stretch the concept of privacy when faced with novel situations. For example, could the use of 'deepfakes' face swapping AI technology (where thousands of pictures of a person are scanned making it possible, via machine learning, to map a person's face onto someone else's with exceptional accuracy) be a "misuse" of private information and not just a breach of the GDPR? In circumstances where this type of technology has been used to accurately superimpose famous female faces onto the faces of actresses in adult films, it is not hard to see how the use of such technology to create this type of content without consent could be a breach of privacy and unlawful.

The GDPR came into effect in May 2018, and provides individuals with a number of rights in relation to the use of their personal data and will also be used to protect privacy in the wider sense. In our view, the GDPR offers additional protection to that afforded under the tort of misuse of private information. First, the definition of "personal data" goes much wider than the definition of "private information", and, secondly, the concept of "processing", is of something ongoing and quite different from a single "disclosure", after which privacy (in terms of the tort) is said to be lost (and with it, the right to injunctive relief). The issue of public disclosure is not relevant to the question of whether personal data is being processed in breach of the GDPR (and an individual's rights).

While some will be more emboldened to enforce their rights; others will be less so going forward. The law of confidence has seen something of a backlash. In what was dubbed 'the UK's #MeToo scandal', Lord Hain used parliamentary privilege to disclose the identity of an individual who obtained a breach of confidence injunction against the Telegraph to prevent the disclosure of damaging allegations which were confidential under a previously agreed NDA. It is not clear whether 2019 will see the 'death of the NDA' (we suspect not), but those entering into them will be far more cautious as to their enforceability in certain situations. Those seeking to enforce contractual obligations of confidence through the courts will have to take not only the law, but also the court of public opinion in the event of disclosure, into account when weighing up whether or not to injunct.

Lastly, the legal methods by which individuals enforce their privacy rights may also change. In Cartier v BT, where Cartier (and others) obtained a blocking injunction against foreign websites selling counterfeit goods, the court made it clear that the basis for granting the order was domestic via the wide discretion extolled upon judges under the Senior Courts Act (1981) (see our previous report). While the EU Enforcement Directive (which also supported the reasoning) does not apply to private or confidential information or personal or special category data (and there is no mention of ISPs in the recent Trade Secrets Directive), there are no such limitations in the Senior Courts Act. This opens up the possibility of the courts blocking access to foreign websites publishing private or confidential information or illegally processing data. This would be a fascinating development, especially given the difficulties with enforcing English court orders in the USA against publishers based there, regardless of the fact that the English courts still have jurisdiction to hear and decide cases brought against them here.

Brexit

It's impossible to think about the coming year without considering the impact of Brexit – if and when it happens. We have written at length on various aspects of the impact on Brexit on IP law (see TMs – No Deal, TMs - Deal, Copyright & Database Rights and Media & Entertainment) but it is also worth considering how UK law might diverge from EU law, post Brexit. Of course, this is only likely to be relevant in 2019 if there is a no deal Brexit as, during a transition period, things will effectively carry on 'as is'. In our view, the most likely areas of divergence are the parallel imports regime and the law of copyright.

Parallel imports

How the UK will address the law of exhaustion of rights in the long-term is of particular concern for those at all levels of the supply chain. At present, once goods are put on the market in the EU by a rights holder or with its consent, the rights holder cannot prevent those goods from circulating within the EU/EEA - its rights are said to have been exhausted.

We know that the current regime will continue until further notice irrespective of whether there is a deal or no deal Brexit. This is provided for in the Withdrawal Agreement and in the draft Intellectual Property (Exhaustion of Rights) (EU Exit) Regulations 2018. However, while the UK government has said that it will retain the current regime, the EU has been silent on the issue. Assuming the EU does not reciprocate, it means that goods first put on the market in the EU/EEA will be able to circulate into the UK. However, goods first put on the market in the UK will not be able to circulate around the rest of the EU/EEA. This will benefit those in the UK who rely on importing goods from the EU/EEA, for example, for sale on the domestic market. However, it will be bad news for many UK-based rights holders and distributors who will have to compete with EU/EEA imports on the domestic UK market but may not be able to sell into the EU/EEA themselves.

Going forward, the UK has said that any substantial changes to the parallel imports regime will occur only after a full research programme and consultation. Rights holders and those in the supply chain are already lobbying for their preferred longer-term positions. It seems unlikely that the UK will continue with the existing regime in the long-term if the EU does not reciprocate for the reasons described above. If the UK adopts a 'fortress UK' regime (where rights holders are able to prevent goods moving from the UK to the EU/EEA (and vice versa) based on their IP rights), then it will be a more level playing field, but policing such a regime with the Irish border will be difficult. It will also adversely impact UK imports and exports. For these reasons, some are arguing for the current regime (EU-wide exhaustion) to remain after Brexit. Whether or not that is feasible – and whether the EU will agree to it – remains to be seen.

In the copyright context, it will also be open to the UK legislature post Brexit to make its own economic assessment as to the types of rights that are capable of exhaustion and to extend the concept beyond the distribution of tangible goods.

Copyright

UK law may also diverge from EU law in the field of copyright. There are a number of areas where (in the absence of any agreement on transitional arrangements) we expect differences in approach to emerge. The treatment of new categories of "work" arising from technological developments is likely to become a more pressing issue in 2019 and beyond (as discussed in more detail below).

Other concepts introduced by the CJEU for the sake of internal market harmonisation may also cease to make sense for the UK after Brexit, for example the idea under Deckmyn that copyright exceptions under the InfoSoc Directive should be interpreted as "autonomous concepts of EU law" rather than under domestic case law. We also expect to see litigation up to the Supreme Court to resolve inconsistencies in pre-Brexit CJEU case law, for example, in relation to the scope of the communication to the public right. Remaining uncertainties in this area include the question of whether rights holder authorisation is needed to retransmit a work by different technical means – this is an important issue for rights holders and distributors as the media industry continues to develop and exploit novel means of content distribution.

The impact of AI

Businesses will almost certainly adopt artificial intelligence (AI) far more next year and in the coming years. We predict that it will become essential for businesses and their lawyers to understand at least the basics of how AI works and how it differs from traditional software. This will be essential to ensuring that IT contracts are properly drafted and any AI generated content is fully protected.

A computer or robot can be said to display or use AI if it does the sorts of things which humans do. While "AI" is frequently referred to, especially in marketing claims, it can mean different things to different people. Two of the most important (real AI) concepts to understand are machine learning (ML) and artificial neural networks (ANN). These can be contrasted with traditional software which essentially comprises a set of deterministic rules. In simplistic terms, ML involves the machine learning by generalising from its experience without being explicitly programmed in the traditional sense; a finite number of examples or inputs are fed in to the machine which learns using probability rather than rules-based software. ANN are vaguely based on the connections in the brain.

AI and commercial contracts

We expect to see businesses (and their lawyers) develop a better understanding of how AI works and, consequently, how to protect ownership. IT contracts relating to the creation or use of traditional rules-based software have largely focused on intellectual property and confidentiality clauses. The value and confidentiality is usually in the source code which may be protected under copyright law. Unauthorised copying may amount to a copyright infringement and/or breach of confidentiality and contract, among other things.

The work in creating ML software can be in the teaching rather than the source code. Therefore, it is possible in some circumstances, for the machine to be replicated or used without infringing any copyright belonging to the original party. If a third party uses the known inputs and outputs to teach its own machine in competition (by getting a free ride on the hard work engaged in teaching the original machine), this may not infringe copyright. A key way to protect ML software from being exploited without consent, eg when it is being licensed to a third party, is through suitable contractual clauses. Ideally, where appropriate, there should be express clauses not only about exactly what uses of the ML software/machine are permitted or prohibited, but also in relation to the use of the inputs and outputs (and anything else which is of value but may not amount to a copyright work or confidential information). A prohibition on the contracting party using the benefit of the ML to teach its own machine may give further protection. We expect to see these types of clauses becoming standard in ML contracts.

AI and copyright

AI is no longer used just to do complex and huge calculations, but is now sufficiently sophisticated that it can be used to create articles, music and art, for example. This gives rise to interesting copyright questions in relation to AI-generated works.

AI-created "works" (such as AI-generated artwork and music) do not easily fit into the "author's own intellectual creation" assessment of subsistence and infringement under CJEU case law. In other words, it is not clear that copyright subsists in works created by AI (where there is no human input) under EU law. That is a significant problem for those looking to invest in AI. After Brexit, the UK may be free to create a new sui generis right for "computer-generated works" under the CDPA, or to revert back to the previous domestic "labour, skill and judgment" test of originality. Under that test, it is much easier to argue that copyright subsists in AI-generated works.

Even if the UK chooses not to diverge (or has no opportunity to) from EU law in this area, we are likely to see questions around the ownership of AI-generated works come before the courts, both at a national and EU level, and this could well lead to a change in EU law.

If you have any questions on this article please contact us.

Regulatory

Privacy & Cookie Policy

Terms of Use

© 2020 Taylor Wessing LLP