Legal’s AI Paradigm Shift—So What?

This blog series by Peter Impey, Consultant General Counsel at InTouch, looks at what AI is, how it’s evolving, and what it means for legal services. It outlines key developments and use cases, touches on risks, regulation and the SRA’s position, and explains InTouch’s practical, low-risk approach to bringing AI into everyday work in a law firm. You can find the full series and related articles on the InTouch blog.

Paradigm shifts occur when there is a profound change in the fundamental model of a subject. As I explored in my previous blog, in my view, the first big shift in legal services was from the ‘Quill to the keyboard’. Now we are witnessing a shift from ‘Keyboard to Cloud’. 

The new normal cloud-based systems focus on data, digital information and scalable virtual resources with self-service by lawyers and SaaS providers enabling networked platforms.

The shift to this environment is accelerating because AI is already cloud-based. LLM powered law firms are becoming part of the new normal and speeding up acceptance of the new normal. 

So what?

Asking “so what?” is a classic way to test a situation for value, relevance, and impact. Taking a step back and asking the question helps you understand the bigger picture and put an issue in context.

The answers bring the focus back to what matters. They can surface hidden risks or clarify a course of action for coping with change. Understanding relative importance is a first step in setting priorities.

Here, the focus is the shift in legal services being driven by generative AI. I’ll consider what we know, and ask “so what?” to draw out the implications.

Knowledge

 We know…

Generative AI reduces the effort needed for knowledge-based, text-heavy work. Models are getting more capable, and cheaper. How do we know? Well…

• Sophistication is up
  The jump in sophistication of large language models is clear. On many benchmarks, models score over 90% and in some cases outperform humans. A newer test, Humanity’s Last Exam, was created to better evaluate performance because many standard tests are now too easy.

• The LLM trend is exponential
  The improvement trend looks like a doubling every seven months, and it could end up being the biggest technology trend of our lifetimes.

• And cost is down
  Progress is accelerating as costs fall. At OpenAI, token costs fell by 90% in a year, even as responses improved. On the hardware side, Nvidia’s DGX1 delivered 1 petaflop in 2016 for $300,000; nine years later, the much smaller DGX Spark is $4,000.

 So what?

Generative AI is highly likely to displace some or even many parts of knowledge work. In that sense, it affects “brain” work in a similar way that industrial mechanisation replaced some “manual” work.

Right now, we’re in the early stages of a productivity surge that could have significant effects on the labour market. For law firms and legal professionals, the challenge is straightforward: the nature of the work is changing.

Competition

 We know…

• The competitive gap compounds: small gains in speed, quality, and cost add up across matters and over time.

• Client expectations shift: what feels “fast” or “good” today can quickly become the baseline.

Unless businesses and individuals adapt to this new reality, they risk falling behind competitors. Finding a sustainable competitive advantage may become a matter of survival.

 So what?

The risk is simple: if your law firm is outperformed by other law firms that use technology better, you may lose work — and over time, lose ground as a business. And if computers can outperform humans at parts of legal work, what work will be left for people to do?

We may be seeing a “Spinning Jenny” moment for office work. Instead of the textile industry, think about anyone reading, analysing or creating text.

In 1764, the amount of work needed to produce cloth fell because of new machines. Now, the amount of work needed to review, manipulate, analyse or produce text falls because of generative AI models.

The argument is that if a company uses automation and LLMs, it needs fewer people. In a widely quoted report, Goldman Sachs suggested in 2023 that generative AI could “expose the equivalent of 300 million full-time jobs to automation”. That would be a major disruption to the global labour market.

Written Work

 We know…

• Legal work is text-heavy by default: most inputs and outputs are written, and often repeatable in format.

• Time goes into handling the text: reading, drafting, summarising, comparing, and extracting key points.

Generative AI is especially relevant to the legal sector because so much of the work is text-based: documents, emails, contracts, legislation, regulation, playbooks, policies, advice notes, board minutes, Articles of Association. It’s all text, and that’s where LLMs can have the biggest impact.

 So what?

An LLM can’t fix a leaking radiator. Plumbers, construction workers and hairdressers are likely to be less affected by generative AI than lawyers. But after office and administrative support, legal work is considered particularly vulnerable, with estimates that around 44% of legal tasks could be impacted.

For most knowledge-based professional services, AI tools are likely to become part of day-to-day work, taking on more routine reading, drafting and analysis — often faster than humans.

Routine Tasks

 We know…

• The tech is already ahead of the bar for many routine tasks: summarising, drafting, and reworking text is often well within its capability.

• Most legal work is about applying judgment in context: not academic depth, but accuracy, risk awareness, and client-specific decisions.

Generative AI tools can already perform at PhD level on some measures. Yet most daily legal work doesn’t require that level of capability to deliver good client service.

 So what?

Do you have a PhD? Do your colleagues? Are you now in competition with the computer?

Most people at a law firm don’t have a PhD. Only around 2% of people in the UK have a doctorate. And most of the day-to-day work involved in delivering legal services simply doesn’t require that level of educational attainment.

The point is to use the technology’s capability in a practical way in the ordinary course of business. That’s the starting point for InTouch, and I’ll come back to this at the end of the blog series with examples.

AI Adoption

 We know…

• The direction of travel is clear: tools keep getting better, and they keep getting cheaper.

• Adoption lags capability: many firms are still testing, experimenting, or holding back on day-to-day use.

Usefulness is improving and so is reliability. While capability is rising, cost is falling. That combination matters, but adoption is still relatively low.

The scale of impact is large, and it’s only just starting. Joseph Briggs notes that only 9% of companies in the US use AI for “regular production for goods and services” in the last few weeks.

 So what?

The impact of AI is still relatively small, but it will clearly build steadily. Moving towards AI is a journey, and it is only just starting. It isn’t too late to begin, but delaying increases the risk.

Law firms that adopt the technology and learn how to use it across client service delivery and business operations will build a sustainable competitive advantage.

Client expectations will change. Smarter use of technology, and the time and cost savings that come with it, will become the new baseline for legal service. Clients will expect better service, faster timelines and improved outcomes. The familiar ask is “more for less”.

Firms that don’t meet expectations will fall behind. They may not even know why the phone stopped ringing. Referrals are a key part of new business for a law firm, and poor experiences or missed expectations do not lead to more referrals.

In this new world, the kind of referral comment a firm needs will be different. The service you provide has to be different too, supported by technology and delivering more value.

The human-to-human relationship will become even more important to the success of your business.

Starting with simple, embedded AI and proven use cases, and building strong partnerships with your technology providers, can speed up the change lawyers need.

 Next in the series, we move from the big picture to real examples of AI in work and everyday life, and what that means for legal service delivery.

This series of blogs about AI is written by Peter Impey, General Counsel for InTouch. If you’re exploring how to use AI in your firm, you can find the full series and related articles on the InTouch blog.