The Coming AI Tsunami
Can you remember what you were doing on October 21, 2015? A seemingly normal day that came and went, but there was something special about that day. This was the day long into the future that Marty McFly in the 1989 film Back to the Future Part II time travelled to.
It’s a strange feeling watching BTTF today, given Marty time travels forward to a date that is now in the past. BTTF’s 2015 had flying cars, hoverboards and robot waiters. The real 2015 didn’t quite live up to it, but I think the film was just early. Films have an uncanny way of predicting the future. In 2011 we watched Contagion, a blockbuster about a worldwide virus killing millions. Just a film and they had Matt Damon to save them, yet less than ten years later we were living through an unprecedented world-wide lockdown.
Now, its AI that is turning sci-fi into reality: I-Robot, Minority Report, Tony Stark’s talking computer in Iron Man and moving photos from Harry Potter are all concepts that AI has made realistic and useable.
AI is not another jump in the advancement of technology like smartphones, tablets or 5G have been. It is gearing up to be a seismic shift for our world.
A tsunami is caused by movement beneath the surface. Before the wave arrives, the waters first recede before they come crashing back in. We have all heard the story of the little girl in the 2004 Boxing Day tsunami who saved people on the beach because she recognised the signs from her school geography lesson.
Like a tsunami, the change that AI will bring in is taking place under the surface right now, but the signs are there to be recognised. It is being silently discovered and adopted by individuals and organisations across the world. A new cold war appearing between America and China and looming war in Taiwan over semi-conductor chips.
The silent movement will change and soon we will see the affects of AI on our everyday lives. I think that change is coming about quickly, or at least the realisation of that coming change. Possibly within eighteen months to two years.
Those who carry on as if nothing is changing will be caught by the wave.
I started Ai Law in 2017 (www.ai-law.com). We are a UK SRA regulated commercial law firm offering the services you would typically expect of such a firm, along with advising on modern areas involving the AI space (not code but online businesses and industry where AI is playing a significant role). Contrary to the name, we are not an AI firm. We are a firm of human solicitors. I see first-hand the ebbs and flows of the legal market and our clients’ businesses, whilst witnessing the current challenges for business owners and law firms around recruitment and managing workloads in our modern fast-paced world. We are actively exploring AI to utilise it within our practice to help improve efficiencies and provide a better service to our clients.
Based on my dealings with AI and experience I can foresee an upheaval in the professional services industry and its current ways of working.
AI is here, but is still yet to arrive. It is beginning to be adopted as a tool to help people to do their jobs, but soon the tool will be able to do the job itself. What will big firms do when this option is available to them? Do they take it, or keep investing in their staff?
Currently there are accepted practices over how services are delivered because that is how its always been. However, AI will be able to do in ten minutes what used to take a day. For productivity this is great, but for a professional service industry built on pricing work based on time and manpower, this is inevitably going to call these current assumptions into question.
AI is capable of significantly reducing the cost of production which will inevitably impact employees who in a professional service business are the largest cost of production.
The cost of labour in a law firm is particularly high. The traditional model of law firm growth is for lawyers to bill x3 their wages: x1 for the lawyer, x1 for the firm, x1 for the profit to the partners. It is no longer this simple in our modern world, as different models have developed within the legal industry, but the concept of targets and x3 base salary remains.
City law firms and American firms in London are renowned for their recruitment drives, each trying to gain an edge. They entice graduates and newly qualified lawyers with eye-watering pay packets on the premise that they can charge them out to clients for their time at a large uplift.
Should there come a time when AI operates a central role in the production of services at these firms, the natural question to ask will be what will happen to the current model of charging when the time it takes to produce the service is substantially reduced. From there the question will be, what will happen to the ability for fee earners to hit their required targets when the output necessary to meet them must also increase.
AI has the ability to become the ultimate worker. Its intelligence is increasing and it isn’t concerned with working time regulations, claiming overtime or bringing employment tribunal claims. At the same time, new employment legislation is being passed by Parliament designed to protect employees by increasing obligations and risk on employers, including higher employer NI and pension contributions and accommodating employees requests for time off, remote working and flexi-time. There is little incentive on employers not to embrace AI in place of human labour to produce its services, if there is a viable AI option available to them. This is a significant risk.
In addition to these risks on the ground, the potential changes raise questions over how to maintain the quality of our professional service industry if AI is fully adopted. If firms opt to utilise AI over employing staff and training them up through experience this could result in an experience drought in around 10 years’ time, which is a serious risk to quality and the UK as a whole. Our judiciary and parliamentary systems have been revered around the world for centuries based upon its history and ability to build on previous law and practices.
People are currently testing AI and seeing its benefits. It can spot typos in a thousand page document within 30 seconds, draft clever sounding legal argument and compute complicated tax calculations. It is obvious why it will be embraced.
In some sense, AI is the ultimate leveller. It allows unequipped people to appear equipped. It lets an inexperienced person go toe-to-toe with somebody who has qualifications, education and experience… or at least its gives them enough to be a nuisance for them. However for the inexperienced person AI can be deceptive. It creates a false sense of security in its perfect prose and sets them up to fail. Litigants in Person are taking to the courts with garbled ChatGPT prepared legal cases, although it should be noted that the quality is improving.
Right now, the idea of using AI to draft emails is considered taboo. Nobody wants to admit to using it. I’ve actually gone to effort to make a point of not using it. It feels wrong somehow, like it’s a naughty secret and I am cheating. But, that perception will change.
The intelligence of the AI models are rapidly increasing. Only recently, since around Christmas 2025, has AI really been intelligent enough to be relied upon in a professional setting. That isn’t to say it can be relied upon fully or copied verbatim, but the quality of responses is much tighter and more relevant than before.
New upgrades are being released faster and the AI is in a race towards super intelligence. Anthropic and OpenAI are the two giants dominating the AI space, both racing towards going public. IPOs are anticipated in 2027 and are expected to pull trillions of investment into them. It seems clear that the way in which organisations are going to spend their money is going to change.
Currently, AI is the new kid on the block being treated as the latest interactive app on our smart phones that we can go to instead of using Google search. We are yet to witness its true power.
For professional services, I believe we are in the stage of research and development. Everybody who is interested by AI and what it can do to improve the running of an organisation is at a similar stage. They are all discovering how it can be embedded into their business. It is when this stage is over and AI is implemented across organisations that we will begin to feel its impact.
It can take longer for larger organisations to adopt new ways of working. There will be ongoing contracts to consider and questions of who to turn to for the right development. No doubt, in typical fashion, the civil service will enter a long and clunky contract costing millions to develop a behind-the-curve AI solution.
For professional services like law or accountancy, regulation and external bodies are a further barrier to implementation. The court system, HM Land Registry, “.gov” websites and MTD are all rather slow in uptake. Even if AI automated every part of a business, that business would still have to deal with other organisations who may not use AI. I like to think of law as a game of tennis. You deal with an issue and put the ball squarely into somebody else’s court for them to return it to you. No matter how efficient you may be, if your opponent is not, the transaction will still be subject to the usual delays.
Since COVID there are many delays and backlogs in government departments today and it can be very frustrating for lawyers and clients alike to progress an instruction. Ideally, AI will tackle these inefficiencies and improve services, but that will require the organisations to actually adopt it and is a long time away.
What we must remember is that AI is incredibly complicated technology. It is so much more than a chatbot or doc production. Whilst some organisations will be more savvy in their use of AI, others will remain unsure.
The common law system is hundreds of years old. With new technology there is always the challenge to apply age-old legal principles to new technology. The application of the law of copyright, online theft, misuse and liability and other areas will have to be developed and it takes time for cases to be tested in court and precedents established. In the meantime, regulators must try to protect the public against misuse of tools that they don’t truly understand.
The SRA’s warning notice on the misuse of AI, published in August 2026, names hallucinations as its first concern: AI tools inventing cases, references and facts that look convincing but have no basis in fact. We have all heard comical stories where this has happened. But, this isn’t necessarily a risk of using AI itself. Having run a law firm, I know that staff hallucinate all the time (… “hallucinating” to mean making a mistake or error, as opposed to some form of substance abuse…).
Trainees have always made mistakes or misinterpreted case law and administrative mistakes can be made. That is how we as humans learn and gain experience. The fact AI can make mistakes is not a risk of AI, it is a risk that the senior partner is not properly checking the work before it is sent out. The risk that the SRA is trying to protect against is relying wholly on AI without checking it, but a lawyer and law firm should never practice in this way anyway and if they did they would be negligent. It is well-established that if you put your name to something, a statement of truth, or cost budget, that you are accountable for its content, no matter who has prepared the work within the firm.
Yet there is a temptation on lawyers relying too much on AI because at the same time client’s are becoming much more demanding, as they are becoming accustomed to instant responses from ChatGPT style chatbots that have the appearance of being humans engaging in conversation, but of course aren’t.
In relation to client confidentiality, the warning notice states, in UK v Secretary of State for the Home Department [2026] UKUT 81 (IAC), the Upper Tribunal commented on the potential for the use of public AI tools to breach confidentiality and legal professional privilege. […] The Tribunal observed at paragraph 21 that ‘to put client letters and decision letters from the Home Office into an open source AI tool, such as ChatGPT, is to place this information on the internet in the public domain’.
But, placing materials in ChatGPT does not publish information on the internet in the public domain and is not open source. Confidential ways of using these tools with proprietary accounts do exist. We certainly cannot be flippant in uploading confidential data and must take the SRA’s warning on board, it is relevant. But, when we input information into ChatGPT, it is handled according to OpenAI’s terms and conditions, which is no different to using a WeTransfer or DropBox or other share drive option that tend to be a commonly accepted practice. These are examples of the gap in understanding and the difficult task regulators and the judiciary have in dealing with something that by the time they are able to pass law or regulation, it is already outdated.
There is a steep learning curve in how to use AI responsibly, whilst at the same time a fast uptake by lawyers and professionals using it to help them work.
I can see the tsunami moment when the R&D stage completes all around the same time and AI is implemented by businesses on a mass scale. It is then we will really understand what this AI is all about. What will be the consequence?
The impact is going to be much more than somebody asking ChatGPT to improve an email or draft a document.
It will be introduced in the private sector first and eventually the public will catch up. It is quite foreseeable that when the private sector finishes its R&D and begins to adopt AI fully, we witness a flurry of new client expectations and demands over the speed of delivery and cost, and the reduction in the size of teams at larger organisations, particularly those jobs which consist of drafting, admin and other tasks that can be replicated and even improved upon by AI.
Speak to Ai Law
If you are interested in anything that has been said in this article, or any legal issues or disputes relating to AI, please contact us today or speak to Tom Ellis.
This does not constitute legal advice and cannot be relied upon. If you need advice please contact us and we will be happy to help.