A recent First-Tier Tribunal (FTT) case was looked at the risks of relying on AI when conducting litigation. The case started when a taxpayer (T), who was acting for herself, appealed a penalty which had been issued against her. In her written submissions, she provided names, dates and summaries of nine apparently supportive FTT decisions, and said that the cases were provided by "a friend in a solicitor's office". The FTT concluded that the authorities were, in fact, fabricated and had been generated by an AI system such as ChatGPT.
When making its decision the FTT gave consideration to a widely publicised US case in which the claimant's lawyers relied on non-existent cases generated by ChatGPT. The FTT accepted T was unaware the cases had been fabricated by AI and did not properly know how to double check these against legitimate sources. As it happened her use of the fictional cases made no difference to her (unsuccessful) appeal. However, the FTT felt that it raised a “serious and important issue" and endorsed comments made in the US case, namely that the submission of fake judgments causes opponents to waste time and money exposing the deception, diverts court time from other important endeavours, potentially harms reputations of judges and courts falsely invoked as the authors of the bogus judgments, and promotes cynicism about the legal profession and judicial system.
Another very recent case related to AI in the context of intellectual property (IP), and in particular patents, and considered the issue of who owns the IP in something “invented” by AI, which would have been capable of obtaining a patent had it been invented by a human. This case went to the Supreme Court, the highest court in the land, who unanimously confirmed a decision of the Court of Appeal that an AI machine could not be regarded as an inventor for the purposes of the Patents Act 1977 and that the owner of the AI machine was not entitled to apply for patents simply because of that ownership.
Another AI related case is proceeding through the court process and heading towards a full hearing. This case also related to IP, but in particular copyright and the law of passing off. The initial claim was brought by Getty Images (US) Inc and related group companies (G) against S, an open-source generative artificial intelligence (AI) company, for copyright, database right and trade mark infringement, and passing off. G is a well-known American visual media company and supplier of stock images, editorial photography, video, and music for business and consumers, with a library of over 477 million assets. G claimed that S had "scraped" millions of images from G's websites without its consent, and used them unlawfully to train and develop its deep-learning AI model Stable Diffusion (SD), and also that SD's output infringed G's copyright.
At a preliminary hearing the court decided that:
- The training and development claim had a real prospect of success. There were reasonable grounds for believing that a fuller investigation into the facts would add to, or alter, the evidence and so affect the outcome of this issue.
- G's copyright infringement claim, which depended on whether the law could also encompass dealings in intangible things (such as making available software on a website), also had a real prospect of success.