The Legal Profession Isn’t Ready For What's Coming: Richard Susskind's Unfiltered Take From Axiom's London Event

Axiom recently hosted an intimate fireside chat in London with Professor Richard Susskind, one of the world's most influential thinkers on the future of legal services and technology. Over the course of the evening, Susskind challenged a room of senior in-house legal leaders to think differently about AI; not just what it can do for them today, but what it will mean for their organizations, their profession, and the broader legal system in the years ahead.

The conversation was wide-ranging, candid, and, at times, deliberately unsettling. That was the point. Below are the key themes and takeaways from the discussion.

Susskind has been thinking and writing about AI and the law for 45 years, longer than most of the audience has been practicing. His most recent book, How to Think About AI: A Guide for the Perplexed, explores the opportunities and risks of AI for society and the future of work. If you want to go deeper into the ideas he raised during the evening, we also sat down with him for a conversation on the Axiom Insights podcast.

The Short Term is Overestimated. The Long Term is Underestimated.

The most important framing Susskind offered was about timing. He believes, and has for some time, that most commentary on AI gets the timeline wrong in both directions simultaneously.

Over the next two to three years, he expects AI to function primarily as an efficiency and productivity tool for legal teams. It will help lawyers work faster, draft better, summarize more, compare more. It will reduce the need for some junior lawyer work. But it will not fundamentally change what legal departments do or how they are structured. The much-discussed transformation of the business model is not coming in the short term.

The long-term story, however, is a very different matter, and most people are dramatically underestimating it. As Susskind sees it, by the 2030s, AI will not primarily be a tool for lawyers. It will be a tool that empowers non-lawyers to handle work that previously required legal experience. Organizations will meet their legal needs in ways that look nothing like the current model of in-house teams and outside counsel. Individual citizens and businesses will have access to sophisticated legal capabilities directly. The very premise of why a legal department exists may shift.

For in-house leaders, this creates a specific planning challenge because you need to be running two tracks at once. The short-term track is operational: Use AI to drive efficiency, use the KPIs you already have, don't overcomplicate it. The long-term track is strategic, and it needs to be happening in parallel, not sequentially. As Susskind put it bluntly, “These are different conversations, and they need to be happening in different rooms.”

The Six Futures of AI

One of the highlights of the evening was Susskind's breakdown of the six competing schools of thought about where AI is headed. He laid these out not to declare a winner but to illustrate why the uncertainty itself demands attention.

The first is what he calls the Hype Hypothesis: AI's limitations will prevent widespread adoption, the bubble will burst, and we will return to our old ways of working. He disagrees with this view, but acknowledges it has serious proponents.

The second is Generative AI Plus: The systems we have today will become fully reliable, human supervision will become less necessary over time, but this is essentially where AI plateaus. What we have now, reliably, is as far as it goes.

The third—and the one Susskind takes most seriously as a strategic planning scenario—is Artificial General Intelligence, or AGI. This is the idea that AI systems will eventually match human cognitive performance across the full range of tasks. Not replicate the way humans think, but match the outcomes and outputs. Everything a lawyer does, a doctor does, an accountant does will be performed at an equivalent standard by a machine. He noted that in 2022, leading AI developers put AGI at 20 to 40 years away. Today, many put it at five to ten years away.

The fourth is superintelligence: AI systems exceed human performance and begin developing the next generation of AI systems themselves, creating a rapid chain reaction of capability. He noted, as one concrete data point, that Anthropic has said nearly 100 percent of the code in recent versions of Claude was written by AI itself.

The fifth is singularity, or the idea that humans and machines merge through digital augmentation of human cognition.

The sixth, which he delivered with characteristic dry wit as the one that ruins the evening, is the replacement hypothesis, or the idea that humanity's role in the cosmos may simply be to create intelligences greater than our own, which eventually supplant us.

His core argument was not that any of these is certain, but that AGI in particular, if it arrives within a decade, would represent the greatest single technological advance in human history. And yet, he said, governments are not preparing for it. Businesses are not preparing for it. The legal profession is certainly not preparing for it. We are, in his words, “sleepwalking.”

The Cultural Challenge is Bigger than the Technology Challenge

Susskind was asked about AI adoption rates. Axiom's own survey data found that in some organizations, as few as 20 percent of legal team members were actually using AI tools that had been deployed for them. He was not surprised.

In his research with his son Daniel Susskind across nine professions, they found that lawyers are the slowest adopters of new technology of any profession except the clergy. That is not where you want to be on that particular league table, he observed.

His prescription is simple and deliberately unglamorous: Every lawyer should be using AI every day. Not necessarily for work. Just using it. Getting comfortable with it. Building a personal intuition for what it can and cannot do. He referenced a general counsel who told their entire team to spend at least 30 minutes a day with the technology for six months, not with a pilot group, not with a structured evaluation framework, just everyone, every day. That organization was transformed.

The mistake, he argued, is the pilot-group approach. The moment you designate 20 people to evaluate AI on behalf of everyone else, you have created two classes of lawyer: those who are developing capability and those who are not. You also end up, a year and a half later, with the non-users still holding strong opinions about a technology they have never actually touched.

He called this irrational rejectionism, dismissing technology without having genuinely engaged with it. It is, he said, something he simply cannot tolerate. More importantly, the market will not tolerate it either.

The legal teams realizing the greatest value from AI are pairing technology with experienced lawyers who know how to put it into practice.

From Reactive to Preventative: A New Role for Legal Teams

One of the most compelling parts of the evening was Susskind's argument about what AI makes possible for in-house legal teams, not in terms of doing existing work faster, but in terms of doing entirely new kinds of work.

He has spent years researching what chief executives actually want from their legal departments. The answer, distilled across dozens of interviews, is simple: Keep me out of trouble. That sounds basic, but it points to a deep frustration with traditional legal service, which has been almost entirely reactive. A problem arises, lawyers respond. Lawyers tell their clients they should have been involved earlier. Nobody disagrees, but the structural conditions don't change.

AI changes those structural conditions. What Susskind described is the possibility of AI agents roaming an organization's data systems continuously, identifying early markers of legal problems before they become disputes, or the equivalent of a blood test that shows early indicators of disease before symptoms appear. Cartel behavior over here. A potential wrongful dismissal there. A defamation risk beginning to surface. A compliance gap. These systems surface problems at a stage where they can still be prevented, not just managed.

He used an analogy he has deployed before: the difference between a fence at the top of the cliff and an ambulance at the bottom. Most of what the legal profession has done with technology and process optimization, he argued, has been building a faster, better-equipped ambulance. AI offers the possibility of building the fence.

This connects to his broader argument about compliance. In many organizations, compliance is still a human-in-the-loop process of checking whether processes conform to rules. What Susskind envisions is compliance embedded directly into operating systems, or rules built into the machine such that non-compliant actions simply cannot be taken. He used the analogy of playing solitaire on your phone and trying to place a red four under a red five, the system does not let you. The rule is enforced by the architecture, not by a person watching.

The Future Lawyer Builds Systems, Not Just Delivers Advice

Running through Susskind's argument is a fundamental claim about what lawyers will actually do in an AI world. He believes the traditional model, or a lawyer delivering advice directly to a client on a one-to-one consultative basis, is not the end state. It’s the current state, shaped by the information systems and human limitations of a pre-AI era.

The future role, as he sees it, is  to build, maintain, supervise, and improve the systems that will deliver legal outcomes. Not to give advice. Designing the infrastructure through which advice, compliance, risk management, and dispute prevention are delivered at scale.

He is particularly interested in what this means for young lawyers. Historically, a new lawyer takes the baton from the person a couple of years ahead and does more or less what that person did. Today, for the first time, there is a genuine opportunity to redesign legal service from the ground up: how legal advice is sought and given, how disputes are resolved, how courts function, how legal risk is managed. The lawyers who build those systems are the ones who will define the profession, or whatever comes after it.

He was deliberately provocative on this last point. He noted that wheelwrights, tallow chandlers, and mercers no longer exist as professions. The demand for wheels, candles, and cloth did not disappear. The modes of production changed entirely. He does not ask what the future of lawyers is, he said, because the question already assumes lawyers have a future, and that assumption may need examining.

Law Firms Face an Existential Question, Not an Efficiency Question

When asked about law firms, Susskind was characteristically direct. He said the question that should be keeping managing partners awake at night is not whether AI will make their lawyers more productive. It is whether a labor-intensive business model is sustainable in a world approaching AGI.

Many law firms, he acknowledged, are already aware of this privately. The near-term pull of AI (more efficiency, more profit, more output) is real and compelling. But the long-term logic points somewhere more uncomfortable. Firms are already reducing graduate intake. The partnership model, built on human labor as the primary source of value, may simply not be viable as AI systems become more capable.

His advice to general counsel on how to engage their law firms on AI is to ask direct questions. What is your view on how this technology will affect legal services? How are you using it now? How are you measuring the gains? And crucially, who is the beneficiary of those efficiency gains? The law firm, or your clients?

He also suggested that in-house teams should not assume law firms will lead the way. Some of the most impressive AI experiments in the legal world, he said, are being driven by in-house legal departments entirely independently. Legal technology used to mean law firm technology. That is no longer true. In-house teams are now building and running systems themselves, and the work that flows to law firms is increasingly shaped by what in-house teams have already done with it.

The Bigger Picture: What the Legal Profession Owes Society

The final thread of the evening was the one Susskind clearly feels most urgently about, and it extends well beyond legal departments and law firms.

He outlined three major societal risks from AI that he believes are not receiving adequate attention. The first is workforce displacement, a hollowing out of white-collar jobs that could produce serious social unrest, particularly among graduates who are already struggling to find the roles they expected. The second is the distortion of democratic processes through AI-generated misinformation and deepfakes, a world in which it becomes genuinely difficult to know what information is reliable. The third is the weaponization of AI in ways that go beyond any previous technology.

On this last point, he was explicit: The closest analogy to AI, in his view, is not the internet, not the smartphone, not social media. It is nuclear technology. Previous enabling technologies brought risks, but not existential ones. This one is different.

He is not optimistic about the pace of governmental response. Regulation is slow, politically compromised, and largely national in  a global problem. He spent time as a special envoy on AI justice for the Commonwealth Secretary General, specifically because he saw the Commonwealth, or 2.6 billion people, including the UK, Australia, Canada, and India, as a possible vehicle for coordinated international influence.

His ask of the room was pointed. Lawyers are lawmakers. Lawyers advise organizations at the highest levels. Aritificial intelligence lawyers shape the regulatory frameworks within which this technology operates. The legal profession should be leading the conversation about AI governance, not as a side note to debates about which platforms to buy, but as a central professional responsibility. That conversation, he said, is not happening nearly enough.

What to Do Now: Susskind’s Three Priorities for In-House Legal Leaders

When pushed for practical guidance, Susskind offered three priorities for any general counsel thinking about AI today.

First, get everyone using the technology. Pick a sensible platform, do not agonize over which one, and make sure it is the whole team, not a pilot group. Usage builds intuition, and intuition builds culture.

Second, build a clear operational plan for the next two to three years focused on efficiency, cost reduction, and productivity gains. Use the KPIs you already have. This is an ops issue at this stage, not a transformation issue.

Third, run a parallel project focused on what the organization needs to become in the 2030s, on the assumption of near-AGI. That journey has to start now. The organizations that are fit for purpose in that decade will be the ones that started thinking about it in this one.