OpenAI CEO Sam Altman says about AI, "We have recently built systems that are smarter than people in many ways and are able to significantly amplify the output of people using them."

In light of his comment, international law firms’ demand for AI has grown exponentially, driven by the success of suppliers such as Legora and Harvey.

Consequently, firms have made investments in AI infrastructure and/or strategic partnerships with AI companies, thus "targeting complex workflows”, as A&O Shearman describes it.

That being said, there has been some doubt cast on AI, particularly regarding client billing and hallucinations. These issues may therefore affect the internal governance of firms, alongside a reshuffling in trainee cohorts.

How are firms investing in AI?

City firms are generally seeking to use third-party AI models to handle more low-value tasks within their workloads.

However, Latham & Watkins recently purchased Nvidia servers for its in-house legal systems for more bespoke reasons. The firm states that the commercial data they handle would not be safe on any other cloud provider. 

Issues with fee structuring

Despite market-wide AI integration, clients have not seen this reflected in their billing.

Most international firms operate under a leverage billing model. This comprises a smaller pool of equity partners and a larger pool of associate staff who handle enormous volumes of work.

This creates longer working hours, where they are billed by the hour, generating high-value invoices, used to structure billing.

However, some of the workload, previously delegated to associates, has been, in part, streamlined with AI, reducing the volume of work. Despite this, clients do not see this reflected in their fees.

Thus, as firms face scepticism over their billable hours, this may raise questions about whether the leverage model is still the industry standard for structuring fees nowadays.

Hallucinations

Not only has billing been called into question, but so has the misuse of AI in court cases, producing hallucinatory citations.

Pinsent Masons and Sullivan & Cromwell have faced scrutiny over AI hallucinations in citations in their court cases. In the case of Pinsent Masons, the firm had to apologise to the SRA for misusing its internal AI pilot. Regarding Sullivan & Cromwell, the firm’s April 9
filing, contained AI-hallucinated citations.

The reality is that AI models show that they are not replacing the role of a lawyer anytime soon, as both cases show how internal models can be prone to hallucinatory citations.

However, as firms increase investment in AI infrastructure, internal oversight will also be necessary to curb staff misuse by staff.

What to look out for

Big law and Magic Circle firms are already making it clear that AI investments will continue; this may affect the following: 

  • As firms reconcile with the Leverage model, we may see a shift from billable hours to outcome-based fee pricing.
  • As firms continue to develop their AI infrastructure, we may see further oversight over its usage in research, such as more frequent AI training or setting up more rigorous governance policies
  • Additionally, the current momentum in trainee cohorts shows a decrease of 2% decrease, primarily due to Solicitor Apprenticeships. This reshuffle of recruitment pathways, along with advances in AI developments, may further affect the number of Training Contracts available.