R (on the application of Officer W80) (Appellant) v Director General of the Independent Office for Police Conduct and others (Respondents)
December 3, 2023
Tech Highlights From Around the Globe – Focusing on Semiconductors and AI
December 5, 2023
GENERATIVE AI: WHAT IS IT?
WHY DO FIRMS USE IT?
WHAT DOES IT MEAN FOR FUTURE LAWYERS?
Generative AI is quickly becoming a popular topic for law firms.
While it had the potential to change the way the law was practised, many future lawyers need clarification on what Generative AI is, what it can do, and why they need to know about it.
What is Generative AI?
Generative AI is artificial intelligence used to create media content, whether text, audio, or images. Thanks to the large AI models that power it, Generative AI can receive, analyse, create, and respond to large quantities of data.
Generative AI is quite sophisticated, meaning it can learn from the data it receives to improve its responses’ accuracy and structure to be more reliable and human-like.
This seems simple enough, but numerous forms of Generative AI have different strengths and include:
Transformer-Based Models
- This G.AI handles sequential data, which is typically good with languages. This can be great for text generation tasks or translations.
Generative Adversarial Networks
- This G.AI has a generator network, which generates an output based on a prompt, whilst a discriminatory network evaluates the authenticity of the output. GANs can create brand-new content or examples based on existing data.
Variational Autoencoders
- Like a GAN, this G.AI runs on two networks: an encoder and a decoder. The encoder simplifies the data that it receives, and then the decoder reconstructs the encoded data into something that resembles the original input. This can also be used to create authentic content, such as a brand-new face.
Multimodal Models
- This G.AI can process multiple forms of data simultaneously.
How do law firms hope to use Generative AI?
The most basic uses of AI are currently present in all industries. AI is used as a customer service provider or performing Q&A tasks. The fact that Generative AI can learn through interactions grants the potential for a vast improvement in how it is utilised for customer service platforms.
Generative AI can also summarise and simplify data, making researching larger quantities of data a much less daunting task. Further, G.AI can be used for repetitive tasks such as translating data or drafting and reviewing contracts and documents.
AI has the potential to save a lawyer lots of time, and time is everything to a lawyer. As a result, many law firms are already incorporating AI into their practices. Allen & Overy, one of the first firms to use Generative AI (Harvey AI), is trained in legal data. Many of the firm’s lawyers directly observe its operation because G.AI still has a long way to go.
AI is known for struggling with the concept of judicial precedent, or the relationship between legal sources. Beyond that, AI continues to make fundamental mistakes of fact and produce slightly inaccurate data. Other firms, such as Clifford Chance’s adoption of Robin AI or Kennedy’s IQ, utilise AI for contract drafting or claim analysis and resolution.
Notably, Kennedy’s IQ is recognised for its innovation in legal AI.
What does it mean for future lawyers?
According to Goldman Sachs, over half of the tasks for lawyers could be replaced by AI in the future. This would change how lawyers operate, are trained, or are hired. Though, as close as that future may seem, AI still has a long way to go before becoming a reliable and accurate operator in the legal sector.
Many law firms have specific aspirations for utilising AI, so any aspiring lawyer dreaming of a particular firm would find it worthwhile to develop their understanding of how that firm uses or intends to use AI.
On that basis, future lawyers must be aware of the potential of AI and its risks.
What risks does Generative AI present in the legal industry?
As mentioned, AI lacks reliability due to the common fundamental errors it can make in fact or accuracy. This is a crucial flaw for AI in the legal industry, as a simple mistake in an unchecked contract could cost clients money and lawyers much time.
A firm can risk misplacing too much trust in AI too soon, eventually damaging the trust between themselves and their clients. However, this is not the only risk that AI can present.
Another risk, and likely the most significant, is AI’s propensity for data security. Using AI in the legal industry means it would interact with the sensitive or personal data of the client. Not only would law firms need to ensure this data is used relatively by AI, but they must also ensure it is sufficiently protected.
This has been a significant international concern in the recent rapid development of AI, especially since the release of ChatGPT. Many clients would likely be apprehensive about trusting AI with their data, so law firms must be as confident in AI as they are in their lawyers.
In the past year, 80% of law firms have had at least one external cyberattack – Allen & Overy being the most recent shock. External cyberattacks impacted client data, and it could be a significant risk to expose such data to artificial intelligence, which could be vulnerable to cyberattacks in the future.
