The challenges of an AI-powered legal industry

[SINGAPORE] Earlier this month, almost 900 lawyers gathered at a Ministry of Law conference to discuss the future of law. From the stage, Law Minister Edwin Tong emanated assured enthusiasm as he spoke of the great strides Singapore’s legal industry has made and how artificial intelligence (AI) presents exciting opportunities.

On the floor – and in the questions sent into a Q&A app – the sense of disquiet was clear.​ How do I retain clients who are already turning to AI for legal answers? How can I use AI while keeping my hourly billing model? Which areas of legal practice are headed for sunset?

How AI will shake up law practice

AI is already changing work for lawyers. Legal publications now offer tools that can generate a legal opinion or research memo in seconds, while other platforms review agreements and conduct due diligence in a fraction of the time a human would need.

Up to 44 per cent of legal tasks can be automated by AI which can do them faster, better. This creates two problems for aspiring lawyers. First, new graduates and junior associates will face increasing unemployment as the work that once sustained them is automated away.

Second, this basic work used to be the classroom in which lawyers learned more complex lawyering skills; without it, where will they acquire experience in critical analysis of legal documents and drafting agreements?​

Law firm owners are facing their own set of existential challenges. Basic legal documents will soon be produced in-house by clients.

The small business owner may get ChatGPT to draft a non-disclosure agreement or a supply contract that is generally serviceable, while institutional clients such as banks will leverage AI to generate run-of-the-mill loan agreements based on standard forms.

Disputes lawyers will not be spared either: AI can review millions of personal injury cases to estimate likely and reasonable outcomes, and simple litigation will be adjudicated by machine.​

Our industry must acknowledge that the entire face of law practice will change. While AI is expected to augment and not replace lawyers, simpler, automatable work will disappear. Instead of delaying the inevitable, it would be better for the profession to pivot.

SEE ALSO

Two lawyers rapped over ‘entirely fictitious’ AI-generated citations submitted to court

Singapore issues guidelines for lawyers on ethical generative AI use

As Tong said, AI cannot build trust with clients. It cannot exercise moral judgement. It cannot advocate with wisdom. This means that the human element in delivering legal services will reside in crafting bespoke commercial solutions, providing strategic advice, and engaging clients with empathy and trust.

Machines will master domain knowledge; lawyers must apply ethics to their advocacy, build trust with their clients and exercise the moral courage to tell a client something they do not want to hear.​

The way I look at it is this: AI is the super-associate who can do much of the legal work. The human role is to be the smart, strategic and effective partner whom the client trusts. But clients will only pay lawyers for more complex or bespoke work, so we must upskill to be a legal product that AI cannot generate or where the human value-add is tangible.​

Enough navel-gazing. What does this mean for the stakeholders?​ Let’s look at different parts of the production line of making a lawyer.​

Law schools

The law curriculum for preparing undergraduates for practice or in-house roles must change. Universities must continue to teach the principles behind areas of law, but academic data-dumping should make way for a more analytical judgement of how the law should be applied.

In other words, less focus on the “what”; and more on the “why” and the “how”. Tong acknowledged that universities have to teach the use of AI in their curricula; I would say that does not go far enough. If people skills, critical (and ethical) analysis and human engagement are crucial, then the equipping must be part scholastic ivory tower, part vocational institute.​

Law firms and training

Within law firms and the wider fraternity, how we train our young lawyers also needs shaking up. If fresh graduates require senior associate skills earlier in their careers, on-the-job training must deliver that.

Mentorship that starts by asking the new graduate: “How would you respond to the letter of demand, mark up the contract, advise the client on the deal structure – and why?” rather than “Watch me and learn” – a Socratic rather than the current osmotic method – would be a good start.​

Small firms and AI as leveller

For small law firm owners, AI is an as yet undiscovered boon. Sole proprietors have long lamented their inability to attract legal talent, yet legal AI solutions can now generate better work product than most newly qualified lawyers.

Agentic AI can summarise outstanding legal issues, schedule meetings and prepare draft e-mail responses, reducing manpower costs for small firms.

In practice, this means that a sole proprietor with the right AI stack can now deliver research, drafting and document review at a speed and depth that once required a team of junior lawyers.

For small and medium-sized enterprises and individuals who cannot afford “Big Law”, this could expand access to timely, competent legal advice instead of closing doors.

Technology here is a great leveller of professional inequality.​ To gain from this, small firms will have to upskill to more complex work and invest in changing professional practices to get the most out of AI. There is even government funding available via various grants and programmes. ​

Business models and value

For all firms, there will be cost savings because manpower spend is the single largest expense.

Clients are already asking for productivity gains to be passed on in fee reductions. As an industry, we need to create billing models based more on value and less on time spent.

A legal opinion that saves a client millions is surely worth more than the lawyer’s hourly rate, and value billing will also force lawyers to be more efficient. This translates more easily in transactional practice, but the inherent lack of predictability in complex litigation is a different matter.

The industry will need to consider if some kind of outcome-based fee is appropriate, without courting the moral hazard of contingency fee arrangements. ​

Getting serious about adoption

For any of this to happen, all of us – and not just junior lawyers or the tech-keen – must get serious about AI adoption. AI will impact the entirety of legal practice, and the lawyer who does not embrace this today is tomorrow’s obsolete practitioner.​

There is a larger moral issue at stake than just making money. Refusing to engage with AI now may affect public access to justice. AI can advance the profession’s core values: providing better-informed advice; faster relief for vulnerable clients; and more time for lawyers to exercise ethical judgement instead of churning documents.

But AI works by trawling databases for historical information. It tells you what is likely, based on precedents, not what should be. There may be in-built biases in content generated that users are unaware of. This is where legal ethics comes in.

Lawyers, with our admittedly often imperfect grasp of ethics, must “hold the ground” in dispensing legal services. In the face of these seismic shifts, our profession has a duty to harness new tools so that we can hold the ethical line – a line that only human beings can draw.

The writer is joint managing partner at TSMP Law Corp