States are increasingly using technology, such as artificial intelligence, to assist with the administration of government. The World Intellectual Property Organization (‘WIPO’) has identified, for example, that artificial intelligence patent applications for computing in government grew by 30% between 2013 and 2016 (see report at page 51).
The administration of intellectual property is one area where technology is increasingly being used. A 2018 WIPO survey found that at least 17 states, including Australia, had started to use artificial intelligence to administer their intellectual property regimes. Such technologies can reduce the cost of administering patent and trademark regimes, and can improve the speed and quality of decisions, such as by reducing errors. However, at least some commercially available tools are ‘not mature enough and cannot be relied upon’ (WIPO, page 46).
At the same time, foreign investors are increasingly using investment treaties to protect their intellectual property rights. Investment treaties often explicitly include intellectual property rights as protected investments, such as Article 9.1 of the CPTPP. Intellectual property rights can also be less directly covered as investments, such as in Article 1139 of the North American Free Trade Agreement (‘NAFTA’). Under the latter, the pharmaceutical company Eli Lilly challenged the invalidation of its patents by Canadian courts in litigation against a generics competitor: Eli Lilly v Canada. Eli Lilly argued, albeit unsuccessfully, that the ‘promise utility doctrine’ in Canadian patent law (see here) defeated its legitimate expectations and therefore breached Article 1105 of NAFTA.
It is foreseeable that these two trends could converge, with a foreign investor bringing a claim against a host State connected with the technological administration of its intellectual property regime. One ground for the challenge could be that the administration did not afford ‘fair and equitable treatment’ (‘FET’) to the investor as occurred in Eli Lilly v Canada. A key component of FET is the foreign investor’s right to the fulfilment of their legitimate expectations, as the Tribunal in Devas v India observed at : ‘[t]here is an overwhelming trend to consider the touchstone of fair and equitable treatment to be found in the legitimate and reasonable expectations of the parties’. (Australian readers may be interested to know that there are differences in the ‘purpose, consequences and underpinnings’ of a legitimate expectation in Australian administrative law and international investment law, which I have previously discussed here.)
In the next section, I consider how such a claim could take shape.
A legitimate expectation of compliance with the regulatory system
The first step is to understand how the expectation could arise. Legitimate expectations have arisen from the legal framework of the host state. For example, in Bilcon v Canada, a majority of the NAFTA Tribunal stated that an American investor could ‘reasonably expect’ that a Canadian government panel undertaking an environmental approval process for a mine ‘would methodically review the potential impacts in light of the core’ question in the relevant Canadian legislation (at ; see also dissent at ). In the Tribunal’s view (at ), there was a breach of Article 1105 of NAFTA, in part because the panel ‘did not carry out its mandate’ under Canadian legislation.
Assuming that the approach of Bilcon is followed, a foreign investor may legitimately expect a trade mark office would methodically review and reject an application by a third party for a trade mark that was identical or similar to the foreign investor’s. Assuming as well that the third party sought a trade mark for similar goods or services. In Australia, the source of that expectation could be s 44 of the Trade Marks Act 1995 (Cth), a reflection of Article 16(1) of the Agreement on Trade-Related Aspects of Intellectual Property Rights. (For a discussion on IP treaties as the basis for a legitimate expectation see here and here.)
The next step is to understand how that expectation could be defeated. A failure by an intellectual property office to reject a trade mark that is identical or similar could defeat the foreign investor’s legitimate expectation. It is possible that in determining to register that second mark, the administering office would have used technology to identify whether there were similarities between the two marks. For example, IP Australia has deployed a tool that draws on artificial intelligence to search for existing trademark images (WIPO, ).
The experience of the Australian corporate regulator, the Australian Securities and Investment Commission, suggests that technological tools looking for similarity may not notice similarities readily evident to a human. When that process is automated, issues can arise. In G C Swinburne and F J McFarlane and ASIC  AATA 602 the business name ‘Melbourne Children’s Psychology Clinic’ had been registered. By operation of a computer program, a new name was subsequently registered, ‘Melbourne Child Psychology’. The registrants of the first name contested the registration of the second name. Unsurprisingly, the Administrative Appeals Tribunal (‘AAT’) found that the names were ‘nearly identical’ within the meaning of the Business Names Registration Act 2011 (Cth) notwithstanding that the computer program did not. It is therefore plausible that errors of this kind, overlooked by a computing system, could result in the incorrect registration of trademarks and give rise to a claim that the government had failed to meet the FET standard.
The final step is to consider counter-arguments to the claim. One such argument could build on the observation of the earlier Tribunal, Saluka v Czech Republic at , which disclaimed the relevance of domestic law: ‘The unlawfulness of a host State’s measures under its own legislation … is neither necessary nor sufficient for a breach of Article 3.1 [fair and equitable treatment] of the [Czech Republic-Netherlands Bilateral Investment] Treaty.’ Read together with Bilcon, this suggests that while the expectation could arise from the host State’s law, failing to comply with that law would not be sufficient for a FET violation. Additionally, as the Tribunal in Saluka observed, the investment treaty should not be interpreted as penalising breaches for which ‘the investor may normally seek redress before the courts of the host State.’ Indeed, the Melbourne Children’s Psychology Clinic found the AAT to be effective, and it was likely far less costly than investment arbitration.
The increasing technologisation of government administration with the concurrent rise in the use of investment treaties by foreign investors to assert their international law rights to fair and equitable treatment of their intellectual property investments could converge. The success of any such challenge would likely depend on the circumstances of the case, and whether the Tribunal favoured or reconciled the Bilcon and Saluka approach to the relevance of domestic law. Could the role of technology inform the Tribunal’s preferred approach?
Alexander Ferguson is a PhD candidate at the King’s College, University of Cambridge looking at international intellectual property law disputes.