Another Legal Geek conference has come and gone, and it was certainly one for the books! Having been with Luminance since it’s beginnings, I’ve had a front-row seat to the legal sector’s digital transformation and what struck me at this year’s conference was how far the industry has come. This is, of course, largely thanks to the impact of artificial intelligence, which is propelling law firms and organisations alike into a new era of business opportunity.

From the many conversations I had with industry leaders and innovators, it’s clear that AI is the buzzword of the moment. But it was interesting to hear so many legal professionals express their concerns about getting caught up in the ‘AI buzz’ and navigating an increasingly complex market. For that reason, I was delighted to kick off day one with a discussion on how organisations can understand, implement and maximise the value of AI. For those who missed it, here’s what I had to say!

Generative AI and the Risk of Hallucinations

Over the past six months, it’s been impossible to escape the chatter about one type of AI in particular: generative AI. ChatGPT is probably the most well-known type of generative AI, which produces conversational, human-like responses to natural language questions. One thing I was keen to express to the audience is that ChatGPT is, first and foremost, a game-changer. It’s incredibly useful for millions of people when it comes to day-to-day tasks like drafting emails, summarise notes, or generating code. But we should exercise caution when using generative AI in a legal context.

As I told my audience, what we need to understand about generative AI applications is that most (if not all of them) are generalists, not specialists. These models have been trained on vast amounts of data – the whole Internet, to be precise! – meaning they’ve been exposed to all types of quality, subject matter and, potentially, bias. As the goal of most generative AI systems is to provide an answer at all costs, a lot of what users might receive is plausible sounding but factually incorrect output. For that reason, it would be difficult to consider ChatGPT and its counterparts legally accurate and impossible to call it ‘legal-grade’ AI.

It’s important to be aware of the issues generative AI could pose when applied to certain tasks that demand specialist legal knowledge. Examples of so-called ‘hallucinations’ are becoming increasingly common, like the recent instance of a New York lawyer citing fictitious case law invented by ChatGPT. Outside of the obvious reasons for concern, in a legal context this is incredibly problematic. So, we’ve run into something of an issue of confidence and reliability here.

But what’s the solution? My answer is legal-grade AI. That’s what Luminance’s legal Large Language Model is – a system trained on legally-verified documents and sophisticated enough to know when it doesn’t know the answer and should pass the work back to a lawyer.

The Benefits of ‘Legal-Grade’ AI

With so many seasoned legal professionals and innovators at Legal Geek, I knew that few of us would need convincing when it came to the benefits of AI for Legal! But one thing I did want to point out is that these benefits are so widespread nowadays, meaning we’re consistently seeing AI lead to a demonstrable return on investment.

However, if I had to narrow down the benefits to just a few points, I’d instantly point to how AI frees up resources and reduces overheads. For example, with your team using AI as their legal co-pilot when reviewing any markups on a contract, they’ll have company precedents and rationale at their fingertips. That means not a minute wasted for you to come back to a salesperson or procurement head on their twenty questions.

Further, AI is critical in helping you prepare your business or client to face any unexpected challenges head on – for example, international sanctions. Having a single source of truth for all your contracts in an AI-powered repository means that you have all the necessary information at your fingertips when you need to answer key business questions quickly.

How to Successfully Implement AI

The prospect of implementing a next-generation technology like AI into your business may seem like a daunting task, especially when you have to cut through the noise and find the right AI for your organisation, whilst securing buy in from other stakeholders. Fortunately, I’ve assisted countless AI implementations, so here are a few insider tips I shared with the Legal Geek audience:

1. Identify the Gaps in Your Tech Stack: ask yourself what’s currently missing in your business? For example, is your current CLM system capable of storing contracts but not analysing them? Or are you missing the first step in the contract lifecycle: contract generation? Complete a thorough assessment, so you know where AI can fill the gaps.

2. Spot True AI: business leaders are now having to navigate a very complex environment where everything seems to be labelled as AI. Check out this blog by Luminance’s Head of Product Specialists, Ella Hislop, to understand common pitfalls in the procurement process and what you should be looking out for!

3. Embrace Evolution: AI will undoubtedly feel new for many members of your team. Change management is a big part of a successful implementation process. Whether it’s hosting internal roundtables, Innovation Days, or simply getting your ‘tech-sceptics’ hands-on with AI, a culture of innovation in your business provides a strong foundation for successful AI implementation.

4. Establishing your AI Roadmap: have a clear understanding AI expansion across the business to allow for clear metrics and goals to be set and met in your AI journey.

5. Try before you buy! Transparency is key. Seek out vendors like Luminance than provide instant value from the get-go and free trials of their technology. If a company is truly providing AI, then you should be able run your documents in the system and experience value from day one.

Interested in experiencing the value of ‘legal-grade’ AI in your enterprise? Book a demo here.