Reading time: four minutes
What's generative AI?
While AI has been a nebulous topic for decades, especially in the last 12 months, there’s been enormous growth in “generative AI”.
“Generative AI is a type of artificial intelligence that uses machine learning algorithms to learn patterns and create new content, such as images, music or text. It works by analysing training data to understand the patterns and then generates new content based on those patterns. Generative AI has numerous applications, including in creative industries, where it can be used to generate new forms of media.”
As seems par for the course for any article about generative AI, the paragraph above was created in a few seconds by one such AI algorithm, from the simple prompt, “please explain generative AI in under 100 words”. This is a basic and fairly undeveloped use of the concept and while generative AI has sparked enormous commercial growth, with several company valuations (eg, OpenAI, StableAI, Jasper) exceeding one billion dollars, it’s also raised significant legal and ethical concerns.
Generative AI tools have garnered huge social media attention. User posts of humorous images, generated from simple text inputs, have overshadowed complaints from artists and creatives about the style theft and intellectual property infringement generative algorithms inherently commit. As with fears around deep fakes and voice cloning, significant technological advancements such as these can entail tremendous legal challenges with regard to IP law. Alongside additional implications for privacy and data protection, depending on what data the algorithms have been trained on. You might be interested to read my article about a viral “cloned” Drake song.
Legal Developments and AI Progression
For example, the US patent and trademark office, overturning a previous decision regarding an AI-generated work, recently declared that copyrightable works require human authorship. While closer to home in the UK, the Supreme Court is currently hearing a case regarding AI’s patentable rights. Perhaps the most well-known development in the realm of generative AI is ChatGPT, which I’m sure you have all heard plenty about this by now!
ChatGPT is a generative pre-trained transformer (GPT) algorithm. GPTs are part of the broader category of large language models (LLMs), which, in a very reductive sense, are good at predicting suitable subsequent words in a sentence. ChatGPT has proven popular and powerful, depending on its application. The current public iteration of ChatGPT, GPT-3.5, came very close to passing the US multistate Bar exam and the recently released GPT-4 passed with flying colours.
The capacity of generative AI seems to only be increasing, with concerns raised around the risk to jobs and entire industries posed by developments in generative AI. Fears should be tempered, however, as there’s already plenty of pushback against generative AI. Springer Nature, the world’s largest academic publisher, recently published guidelines on AI-authored research papers and another AI algorithm is currently being sued for holding itself out as a lawyer. Additionally, regulatory developments will likely be sought to curb some of the potentially negative impacts of AI.
Rather than fearing AI, junior lawyers and those aspiring to practice would instead do well to cultivate a working knowledge of relevant AI tools and focus on how they can leverage generative AI to optimise their work. Many law firms have already implemented AI features into their workflow, to assist their lawyers in monotonous or routine tasks that can take away from fee-earning work. Harvey is a generative AI tool and while it’s based on GPT-4, it’s a very different application to something like ChatGPT as it’s tailored specifically for legal work. For example, Allen & Overy LLP have recently entered an exclusive launch partnership with the software, handling forty thousand queries over three months, it has enabled the firm’s lawyers to deliver, “faster, smarter and more cost-effective solutions to their clients”.
Generative AI and AI more broadly should be approached with pragmatic caution in the coming years. Lawyers and firms should carefully consider the implications of generative AI tools and keep them aware of relevant legal developments. On a more optimistic note, while generative AI may potentially pose a risk to certain aspects of work, including the work of lawyers, it also presents opportunities to increase efficiency, allowing time and effort to be directed to the successful handling of more significant work. Those that accustom themselves to these technological developments now will be setting themselves up well for a future in law. Law is looking to run with AI for the long term, so aspiring lawyers should get ahead of the curve and get clued up on AI.