Last week, I kicked off our generative AI series with a webinar, compèred by the talented Nigel Rea. The webinar was our highest attended to date (by some margin!) reaffirming that in-house lawyers have a thirst for knowledge and insight into the topic.
This has led me to write this 3-part series on generative AI, designed as a guide for in-house lawyers. I’ll be doing this in my typical straightforward ‘to-the-point’ Yorkshire style, focussing on outcomes and value and providing some candid opinions. Additionally, I’ll be addressing the questions that many of our customers frequently ask.
In this 3-part blog series, I’ll be covering:
So let’s get started…
Generative AI is a subset of Artificial Intelligence technology, which is designed to generate new content (such as words, images, music) based on inputs. It uses a range of techniques to analyse and understand input data, and then uses that to generate new content. Generative AI is built on ‘large language models’ which basically means that it has been trained on a big set of data, on which it has learned underlying patterns in that data, which it then uses to generate new content.
In practical terms, imagine having a colleague or intern, who acts as your personal intermediary, writing an initial advice note based on research they have done, drafting emails based on a set of meeting notes from yesterday, summarising a paragraph into a few words, or changing the style of writing to another “voice” (e.g. a layperson to remove legalese) with the perfect tone to help your recipient.
Generative AI is certainly the tech of the hour. I have no doubt that it will be transformative to the way we work in the coming years. While I’m not convinced that it will quite be to the extent many are claiming, I think it will be a pervasive transformation, seamlessly integrated within many tools and processes we work with every day. We are already seeing that in the right scenarios and setting it adds a lot of value and is really powerful – I’ll talk about that a lot more in blog 2. To gauge though, Goldman Sachs predicts breakthroughs in AI that could drive a 7% increase in global GDP!
The underlying principles and foundations of large language models have actually been around for decades. Recent technology and methodology advances have made the technology much more practical though – allowing us to train with and process much bigger data sets, and doing it much more quickly.
OpenAI brought generative AI to the mainstream through its ChatGPT product. It’s one of the popular tools you can go experiment with for free (just don’t use sensitive data!) and I’d really encourage you to go do so. It’s rare for such powerful new technology to be accessible so freely in the public domain which has helped to fuel the popularity – even my gran has heard of it! There’s no doubt that Microsoft’s $10 billion investment into OpenAI has probably helped with this too. It’s important to mention though, that there are lots of large language models and generative AI tools available. I won’t go into the detail here, but it’s worth a search around what Microsoft, Google and Apple are doing, as well as other ‘open source’ tools.
Not to forget, GPT-4, OpenAI’s latest language model, passed the US Bar exam amongst the top 10% of scores! That’s really impressive!
Legal is heavily touted as one of the major areas ripe for improvement through generative AI. That’s because we are essentially a knowledge and content sector, heavily focused on the written word. We’ve come a long way in the last 15 years or so in terms of streamlining and driving efficiency in legal process but there is so much more to do.
Using technology that can generate content, re-write existing content, help facilitate more sophisticated search and summarise complex language into more simple terms is obviously going to be very relevant to the legal sector.
Generative AI is already being deployed in tools used every day by lawyers, including Microsoft Office, CLM tools and knowledge/research platforms like Thomson Reuters’ Practical Law. Many of our customers and in-house lawyers I’m speaking to in the market are experimenting with Generative AI in a work context to solve everyday challenges. An ever-increasing number have even rolled out use cases into full ‘production’ use and they are demonstrating significant efficiency gains and generating value in other ways.
It’s important to note though, that Generative AI isn’t going to eat lawyer’s jobs. An ever-increasing mass of data being generated in organisations means lots more contracts and data for lawyers to create and consume/review every day. Without this type of technology being developed alongside this ever-growing mass of data, how will we ever get through it all? It’s also helping to cut down on everyday admin tasks which are time-consuming and burdensome.
So, the closing comment here is, Generative AI is a really interesting technology and is already demonstrating its place in the stack of tools and processes delivering value in legal. In Part 2, I’ll talk more about tangible use cases and benefits (plus address some of the challenges and risks) to help bring things to life.
What a great day to share this article...