If ChatGPT is everyone’s personal lawyer - without the billable hours - now what?

A recent post on LinkedIn tells the story of a frustrated SUV owner who used ChatGPT to fight a $10,000 repair bill. He'd been ignored for years, dismissed and delayed by corporate bureaucracy - until he fed AI the relevant consumer laws, warranty terms and past legal cases related to his claim. 

Then he prompted it: "You're one of Australia's top consumer lawyers. You've never lost. Draft me an email that ruins their Monday." Two days later, the car manufacturer replied. They would waive the cost of repair. Then, the clincher. He goes on to propose: "You don't need a lawyer. You need better prompts…"

Regardless of how accurate it might be, this post strikes a nerve within the legal profession because it exposes a reality we can no longer ignore: AI is here and everyone (I mean everyone!) is using it. But for those who actually have legal qualifications, the reality of tapping into AI's potential without compromising the integrity and quality of their legal expertise is less straightforward. 

Jamie Ng, Global Clients & Markets Partner at Ashurst, delves into 'the jagged frontier' of AI in which ubiquitous knowledge is fundamentally changing the way the profession operates and the need to 'hyperscale' insight, judgement and experience in order to remain competitive.

The jagged frontier of AI in law

The undercurrent of stories like this is that legal knowledge is now ubiquitous. Yesterday, lawyers used to pride themselves on their recall of legal precedent, their ability to conduct legal research across a multitude of sources, and even knowledge of the intricacies of different laws. Today, this knowledge is almost instantly accessible across millions of cases, textbooks, journals, legislation and regulation and more to billions of people. Long gone are the days of owning a library of dusty yellowing books or even accessing specific databases on particular laws. 

Unlike traditional research methods which can require painstaking hours of digging through materials, AI can process, analyse and synthesise information at speeds impossible for human lawyers to match. Legal knowledge is now ubiquitous. But while AI can generate legal content, can it apply judgment, contextual understanding, ethical reasoning and strategic insight (all qualities that define top-tier legal professionals)?

Between November 2023 and March 2024, Ashurst, led by my colleague and fellow partner Hilary Goodier, tested AI in legal workflows with 411 staff across 23 offices. The goal was to cut through the hype and assess real-world impact. AI tools saved up to 80% of time on UK corporate filings and 59% on research reports, but accuracy was inconsistent - scoring 1 to 4 out of 5 for AI drafts, versus 3 to 4 for human work. Lawyers found AI helpful to beat 'blank page syndrome' but noted it could create extra work when wrong. In other cases, AI matched or even beat human output  and expert reviewers misidentified 50% of AI-generated content as human-produced; yet flagged the differences in tone structure and depth of legal analysis.

It would seem that the adoption of AI is far from seamless; and that as a profession, we find ourselves at the 'jagged frontier' where its capabilities are effective in some areas but unreliable in others.

The need to hyperscale human insight 

Is the next logical step, then, for law firms to supercharge their AI models with the cumulative centuries of legal insight, judgment and experience held by senior partners? By training AI on the unique knowledge and nuanced decision-making of their most seasoned practitioners, firms can move beyond generic outputs to something far more powerful: the seamless integration of their ubiquitous knowledge with human expertise. 

Using this principle, AI doesn't just draw on what is generally available to everyone but also synthesises firm-specific insights, previous strategies and outcomes and the collective wisdom of top legal minds to deliver faster, smarter and more trustworthy outcomes. It isn't an off-the-shelf AI anyone can download - it's the result of deliberate investment in firm-trained models, grounded in the lived experience, judgement and nuance that only seasoned legal professionals can provide.

Importantly, upgrading the use of AI from straight-forward access to ubiquitous 'knowledge; to more specialised firm-specific 'insight' means that AI becomes a true thought partner for firms focused on levelling up the value they provide their clients at every and all levels of the firm. Better inputs into AI models makes it easier for lawyers to breathe life into its outputs as they pioneer providing a consistently higher level of quality and strategic advisory - shifting AI from being a simple research tool to an active collaborator in legal problem-solving. 

Disrupting the age-old apprentice model

This shift - from AI as a knowledge tool to a true strategic partner - marks more than just a technological upgrade; it signals a reimagining of how legal expertise is built and shared.

Traditionally, legal expertise has been built through an apprenticeship model involving years of mentorship, incremental experience and gradual skill-building. Now, rather than spending years manually acquiring the intricacies of legal research, associates could leverage AI to rapidly understand case law, industry trends and regulatory shifts - meaning junior lawyers can contribute meaningfully far sooner and increase efficiency across the firm with real-time, firm-specific insights at scale. Leveraging the principles of hyperscale, firms also have an opportunity to fast-track access to knowledge traditionally held at senior partner level. Instead of requiring years of experience to develop deep legal acumen, junior lawyers could access firm-specific legal insights from the most senior partners instantly and use this to help breathe life and real insight into the outputs from generative AI. 

This doesn't mean the apprenticeship model disappears, but it does mean that firms have the opportunity to rethink how they develop, scale and deploy legal talent. 

Strengthening strategic offerings

By the same token, hyperscaling how firms access the insights that come with deep experience could help ensure consistency in service delivery, reducing the variability in legal analysis across different offices and teams. This is particularly valuable in complex legal domains where experience and intuition play a critical role in shaping the best course of action. 

Brand-new challenges

Of course, no frontier is without its pitfalls and embracing AI brings with it an entirely new set of risks - including the subtle (but dangerous) phenomenon of hallucinations demonstrated by the now-infamous 2023 US case where a lawyer filed court documents that cited half a dozen fictitious cases generated by ChatGPT. The model had created entirely fabricated precedents complete with plausible-sounding names and citations. The lawyer, unaware of the erroneous outputs, submitted them in a federal court where the judge sanctioned the lawyer and his firm and the story went viral as an instant warning to every legal professional using generative AI as a research tool.

These examples are not just cautionary tales; they're flashing red lights reminding us that legal work doesn't just rely on information; it depends on accuracy, accountability and integrity and as firms explore the idea of training models on past transactions, legal opinions and internal strategy documents, ethical questions will inevitably follow. These will include 'How do we preserve confidentiality?' and 'What are the implications for legal professional privilege?' as well as 'How do we ensure clients' data isn't being inadvertently used to train tools that could later generate outputs for competitors?'

These aren't reasons to stall innovation; but they are reasons to be deliberate and pursue responsible adoption with robust governance frameworks, clear boundaries on data usage and human oversight built into every layer of the AI lifecycle. 

Now is the time for the legal profession to think outside the square. AI is not a silver bullet, nor is it a passing trend. It is the next evolution of legal work - one that will reward firms that adapt and challenge those that resist.