Generative AI

And the Legal Profession

Have you tried the new Instagram filter?

You can take a picture of yourself – sans makeup (for those of us that wear makeup) – hair all over the place, eyes still full of sleep and it transforms us into a glammed up, styled to the nines, Queen of the Runway – Yaaaaaas Queeen!!! Ready to slay the day!

Ok. Maybe not.

But don’t you sometimes wish there was filter that you could apply to your overtired, overworked face that would make you look (and hopefully feel) fresh and polished? If so, you would be referring to Instagram’s variety of AI story options, including generative AI stickers, backgrounds, and filters.

Essentially generative AI – at least in this instance – is a computer model that learns patterns about information, so that people can use it to get answers to questions or create something new like text or images (like a fresh, polished face).

What is generative AI?

Generative AI refers to deep-learning models that can take raw data — say, all of Wikipedia or the collected works of Van Gogh — and “learn” to generate statistically probable outputs – like high-quality text, images, and other content based on the data they were trained on – when prompted to do so.

At a high level, generative models learn from training data and then recreate new work that’s similar, but not identical, to the original data. While generative AI has been used for years in statistics to analyse numerical data, it has been the rise of deep learning, that has made it possible to extend generative AI to images, speech, and other complex data types, like music and paintings.

But the focus on generative AI really came with ChatGPT (by OpenAI) in 2022. OpenAI’s chatbot, powered by its latest large language model, can draft poems, tell jokes, and churn out essays that look like a human created them. Prompt ChatGPT with a few words and out will come the greatest love poems or song lyrics for hit songs.

ChatGPT saw a giant leap forward in natural language processing. And it has seized the imagination of just about everyone – even the lawyers. And it’s not just language: generative AI models can also learn the grammar of software code, molecules, natural images, and a variety of other data types.

It’s an exciting world with generative AI and where it can go is endless.

New add on to ChatGPT!

“Deep research” is a tool that has just been released with ChatGPT (it also already exists through a platform called “Perplexity AI“). Deep research is the newest add-on to the vast capabilities of ChatGPT. Deep research allows you to scour thousands of resources and databases with Perplexity AI or ChatGPT and pull information specific to a topic –

“Tools like (perplexity.ai), which leverage deep research, can assist with tasks such as:

• Answering complex questions
• Generating high-quality content
• Summarising long documents
• Providing insights and analysis
• Supporting research and development

These platforms have the potential to revolutionise the way we conduct research, generate content, and make decisions, by providing access to vast amounts of knowledge and insights.”

Generative AI and the lawyer

Since ChatGPT arrived on the scene, machine learning has demonstrated an impact in a number of industries, accomplishing things like medical imaging analysis and high-resolution weather forecasts. And it’s clear that generative AI can do a lot more than just draft poems or sheets of music.

In a McKinsey & Co survey completed in 2024 the following was set out –

“If 2023 was the year the world discovered generative AI (gen AI), 2024 is the year organisations truly began using – and deriving business value from – this new technology. In the latest McKinsey Global Survey on AI, 65% of respondents report that their organisations are regularly using gen AI, nearly double the percentage from our previous survey just ten months ago.

Interest in generative AI has also brightened the spotlight on a broader set of AI capabilities. For the past six years, AI adoption by respondents’ organizations has hovered at about 50%. This year, the survey finds that adoption has jumped to 72%. And the interest is truly global in scope.”

And the biggest increase in adoption can be found in professional services.

According to a survey conducted by Lexis Nexis that interviewed 1,176 US lawyers, 1,239 law students, and 1,765 general consumers it was found that the legal market is significantly more aware of generative AI than the general population (88% vs. 57%). In fact, 39% of lawyers, 46% of law students and 45% of consumers agree that generative AI tools could significantly transform the practice of law.

How you may ask? Well because generative AI has the ability to perform cognitive tasks in natural language there’s a profound implication for the legal profession, which – quite obviously – relies heavily on language for tasks such as drafting documents, conducting legal research and providing complex advice.

And because generative AI can do this, it has the very real potential of completely transforming (in a good way) the legal landscape. It could even offer those that have previously been unable to afford it, access to justice.

How generative AI can help lawyers

Look, before we blow this whole generative AI thing out of proportion let’s get one thing straight – lawyers already have AI tools at their disposal. Tools like document review software, e-discovery tools, and predictive coding applications. Tools that can help you generate reports, tools that can help you generate documents – such as that offered by XpressDox – helping you to streamline template creation and document generation.

The only difference with generative AI is the latter’s ability to understand and generate natural language, akin to a junior lawyer. Consequently, generative AI tools can perform a wider range of tasks more efficiently and accurately than their predecessors.

Here are three examples of the enhanced capabilities of generative AI –

1. Document drafting and review – generative AI can significantly enhance the drafting and reviewing of legal documents, tasks that all legal practitioners know, involves repetitive, labour-intensive work that’s also required to be perfect. Tools like XpressDox can produce initial drafts based on predefined templates and specific client requirements, allowing lawyers to concentrate on the more complex and strategic aspects of contracts.

2. Legal research – a time-consuming process that often involves the review of vast amounts of case law, statutes, and various amount of legal literature. While search engines can speed things up, they use keyword searches which often produce irrelevant, completely incorrect, or insufficient results. Generative AI can streamline this process by quickly analysing copious amounts of data, identifying relevant precedents, and summarising the key points. This doesn’t just save time but also ensures that legal professionals have access to the most pertinent information, thereby enhancing the quality of the legal advice given.

3. Advisory – AI can produce complex and nuanced advice and arguments in the same way as an experienced (albeit junior) lawyer does. You can’t find these types of arguments in literature and existing legal materials, leaving lawyers to rely on their own expertise and insights to construct innovative and convincing arguments. One approach to transforming generative AI into a true legal expert is through custom training, also known as fine-tuning. Fine-tuning a general AI programme into a legal expert involves several key steps: identifying specific legal tasks; collecting extensive domain-specific data; and customising the model to manage complex legal reasoning. This process ensures that the AI programme can understand and generate relevant, reliable legal content. For example, Harvey, a generative AI platform, has partnered with OpenAI to create a custom-trained case law AI model. This has allowed Harvey to deliver AI systems that help with tasks requiring complex reasoning, extensive domain knowledge, and capabilities beyond a single model call—such as drafting documents, answering questions about complex litigation scenarios, and identifying material discrepancies between hundreds of contracts. During the training process, Harvey and OpenAI collected vast amounts of case law data and refined the model’s ability to perform tasks such as analysing complex litigation scenarios. Harvey worked with 10 of the largest US law firms. They provided attorneys with side-by-sides of the output from the custom case law model, versus the output from GPT‑4 for the same question. By doing so, Harvey enhanced the AI’s proficiency, resulting in outputs that legal professionals overwhelmingly preferred for their accuracy and depth.

What are the challenges of generative AI?

One of the biggest issues with generative AI is the occurrence of hallucinations – which is kind of exactly as it sounds. As described by IBM –

AI hallucination is a phenomenon wherein a large language model (LLM)—often a generative AI chatbot or computer vision tool—perceives patterns or objects that are non-existent or imperceptible to human observers, creating outputs that are nonsensical or altogether inaccurate.

In other words, generative AI models produce inaccurate, misleading, or nonsensical results, like text that appears credible but turns out to be factually incorrect.

In the legal world this is quite clearly a no-no and can be highly problematic for a lawyer where the facts are everything. This was the case against the Colombian airline Avianca in 2023 when a US attorney filed a legal brief that he had written with the help of ChatGPT. The document included citations to six legal cases that seemingly offered precedents that supported his client’s position, but these legal matters didn’t actually exist. They had been invented by ChatGPT. The Judge fined the two lawyers, Steven Schwartz and Peter LoDuca, as well as their law firm Levidow, Levidow & Oberman $5,000.

The judge P Kevin Castel said in a written opinion that there was nothing “inherently improper” about using artificial intelligence for assisting in legal work, but lawyers had to ensure their filings were accurate.

A notable piece of advice.

Incompleteness is another issue. Here the responses produced by generative AI fail to address the user’s query properly or they provide improper case citations. In this scenario, the responses will either be of no use or require the user to add citations manually before they can be used in court.

Another problem with generative AI is inconsistency. Generative AI models produce responses based on the random nature of their training data, which can lead to variability in answers, especially for ambiguous or complex queries. This can then lead to confusion by the users leading to mistrust of the generative AI’s capabilities. Not ideal.

Finally, jurisdiction is a problem. Law – as we all know – is jurisdiction-specific, meaning that each country has its own set of legislation and case law. Because of this, AI trained on the laws in one jurisdiction will not be useful in another. However, generative AI in the financial sector or in the medical sector, trained on the data in one country can be easily deployed elsewhere. It’s the jurisdiction specific generative AI that becomes a problem.

The opportunity for businesses is clear. Generative AI tools can produce a wide variety of credible writing, coding, music, art – you name the data, the generative AI can produce the output – in seconds, then respond to criticism to make the output more fit for the users’ particular purpose.

This has implications for a wide variety of industries, most notably the legal industry that has seen basic awareness of generative AI increase by 86%, where half (51%) the number of lawyers surveyed (1,176 lawyers and 1,239 law students in the United States) have either already used generative AI in their work or were planning on doing so. 84% believe generative AI tools will increase the efficiency of lawyers, paralegals, or law clerks, 61% of lawyers and 44% of law students also believe generative AI will change law schools and the way law is taught and studied. That is significant.

It’s clear that generative AI has made an impact on the legal industry and as Judge Castel said in the Colombian airline Avianca matter – there’s nothing improper about using artificial intelligence for assisting in legal work, but lawyers must ensure that their filings, that their advice, that their research, that their document draftings are accurate, that they are correct and that they can be relied on by their clients.

Simple as that.

In the meantime, if you are ready to incorporate a new tool into your existing accounting and practice management suite, or if you need to start from scratch, Get in touch with us to see how we can best support your strategic goal setting and progress tracking needs. We have the right combination of systems, resources, and business partnerships to assist you with incorporating supportive legal technology into your practice. Effortlessly.

AJS is always here to help you, wherever and whenever possible!

– Written by Alicia Koch on behalf of AJS

(Sources used and to whom we owe thanks: Lexis Nexis; S&P Global; Economics Observatory; OpenAI; The Guardian; McKinsey & Co here and here; IBM here and here and Plann That)