What will The Lawyer’s AI headlines read in three years? This question was posed to managing partners at the Lexis Nexis roundtable, hosted in conjunction with The Lawyer, last month. Since ChatGPT was launched in November 2022, discourse surrounding generative AI has accelerated at a staggering rate.

There is a spectrum of views across the profession, but sentiment is shifting from fear to excitement, albeit on the cautious side. Ultimately, some will be faster to adopt AI than others. The key takeaway from the debate is that AI will create different skillsets and subsequent value proposition. Impacts to firm structure was also highlighted.

Let’s look at some potential AI headlines for The Lawyer 2026…

The 5,000-an-hour hourly rate

Much of the work conducted by junior lawyers, especially at trainee and NQ level, could be done by generative AI in the next three years. Leaving document reviewing and contract drafting to AI allows trainees to complete work of higher value. As a result, young lawyers could spend more time building client relationships.

Trainees charge high rates per hour to complete low value tasks. Adding value through AI could enable junior lawyers to price their work on value instead of time, thereby replacing the hourly rate.

Training contracts slaughtered

Bringing in generative AI to do trainee work means fewer training contracts may be offered, as there will be less work for trainees to usefully do. Fewer training contracts could subsequently alter a firm’s hiring structure from a pyramid shape, with trainees at the bottom and partners at the top, to a column, with little differentiation between the numbers of trainees and partners.

Consequently, junior lawyers would have to acquire key solicitor skills in a different way to conducting lower value work as a trainee. Roles acting in a similar manner to pupillages, where juniors shadow more senior lawyers, could emerge. However, this creates equality, diversity and inclusion issues as most of the population cannot afford to shadow for low pay, unlike trainees with permanent contracts. Barriers to enter the solicitor profession could therefore increase.

Changes to trainee workloads will ultimately dictate training needs. Altering the structure of the firm and innovating to prevent unwanted changes caused by AI will cost money. Therefore, not all the cost efficiencies created by AI will be to be passed onto the client. This doesn’t bode well for clients, who love to know what alternative fee structures are available.

Lawyers out, data scientists in

Ultimately, lawyers could end up making less money, with data scientists’ salaries rising. Wider education trends show less impetus on humanities subjects versus STEM… maths is now the most popular A Level. Top AI specialists are charging high fees for their work. Will firms be able to attract and retain top tech talent when their salaries are higher than that of partners?

Equally, there was debate on the extent to which AI will make judgements. Some argued that it is unlikely that AI will go that far. Lawyers earn their fee by giving advice through which they make a prediction and judgement. Effective prediction and judgements are determined by a multitude of human factors, such as personality, expertise, and experience. The importance of conducting background work to reach a judgement cannot be underestimated, as one attendee noted that writing forces you to build an argument by weighing up points of view. Furthermore, the roundtable agreed that making judgements is cathartic, so it is unlikely that lawyers will let AI play a significant role in the court process.

However, others suggested that advice will become more data driven with the use of Gen AI. Data can provide immediate context by way of predictive analytics. Some also see AI judgements in our future, where two parties agree to a form of AI arbitration with an agreed dataset and tech stack.

Infrastructure over individual?

If datasets become an integral part of how arbitration operates, clients may choose firms based on their technology stacks rather than prioritising reputation and expertise. Whilst we may see clients focusing more on infrastructure than the individual, it is important to note that human input is still crucial in litigation and disputes.

AI could change how the legal sector operates in-house, as teams could find themselves outsourcing less work to firms. AI would allow them to draft contracts in-house, then send to law firms for a risk assessment, and they could feed the contract through AI again.

Consultancies could also emerge on how in-housers can use AI. The independent legal profession would shrink. Therefore, firms must constantly regenerate and innovate. Ultimately, challenges remain in encouraging the profession to view Gen AI as an opportunity to use technology to speed up what lawyers do for clients.