FutureProof: The rising use of AI by the legal profession

Compared with some professions, lawyers tend to be slower adopters of new technology. However, the percentage of lawyers using Gen AI is increasing. In the second of our new FutureProof series, we look at how lawyers are adopting Gen AI and the opportunities and risks which this presents, both for lawyers and their professional indemnity insurers. 

Fast development of Gen AI use and adoption

In September 2024, LexisNexis undertook a survey of over 800 law firms regarding their use of Gen AI. That followed similar surveys in July 2023 and January 2024. The results show a significant increase in Gen AI usage, with those saying they used it at least once a month rising from 11% in 2023 to 41% just over a year later. On top of that, a further 41% of those surveyed said they were planning to use Gen AI in the future and only 15% of law firms surveyed said they had no current plans to adopt AI. Read the report on the survey.

The trends highlighted by LexisNexis’ survey are not surprising given that Gen AI has become a global hot topic, not just for lawyers but also for their clients. Businesses across most sectors are looking at how Gen AI can give them a competitive edge and generate cost efficiencies. Inevitably, they will expect their legal advisors to do the same.  

Who is using it? 

As is often the case, larger firms have generally led the way with their adoption of Gen AI.  

Research indicates that in 2023 many larger law firms significantly increased their “research and development” spend, concentrating on Gen AI. Some firms have developed their own Gen AI products.   Others are integrating externally developed products such as Harvey, Co-Counsel and Microsoft Co-Pilot. Towards the end of 2023, and into 2024, we have seen those law firms launch products and begin to embed the technology into their practices.  

We expect that trend to continue. However, the adoption of Gen AI isn’t limited to larger firms. Many mid-tier and smaller firms are piloting Gen AI products and, judging by LexisNexis’ recent survey, the day-to-day use of Gen AI is growing exponentially.   

This has already been recognised by the  SRA, which was quick off the mark to produce a Risk Outlook Report (November 2023) on the use of AI by the solicitors profession.

LexisNexis’s survey also indicates that in-house lawyers are embracing Gen AI on a faster scale than other lawyers. Barristers are also making the leap towards it, although to a lesser extent than their solicitor counterparts. Indeed, the Master of the Rolls (Sir Geoffrey Vos) recently expressed his view that the adoption of AI is something that all lawyers, without exception, need to understand and consider. Read his speech on the issue.

How are lawyers using Gen AI?

There are conceivably hundreds of “use cases” for GenAI in the legal market, each carrying a different level of benefit, difficulty in implementation, cost, and risk (amongst other things). Goldman Sachs predict that 44% of legal tasks could be automated by AI in the future. 

Whilst many lawyers have already been using some form of basic AI and machine learning for some years to assist with various legal tasks (such as e-disclosure), the more intuitive and generative nature of GenAI opens up much greater opportunities. For instance:

• Whilst a lawyer might have used basic AI previously to scan a document for key terms, Gen AI can scan hundreds of contracts (for example) and provide a summary of their contents, identifying terms of concern and even recommending amendments.
• It can also go a step further and draft documents such as agreements, legal submissions, correspondence and even questions for cross-examination.
• Previously, lawyers could use basic AI to identify useful case law. Gen AI takes this a stage further by identifying a range of relevant case law, summarising  legal precedents and providing text for legal argument by reference to that case law.
• Gen AI can also be used to analyse data in relation to previous cases in order to predict court outcomes and recommend strategy. There are already examples of lawyers using Gen AI to help identify new legal theories and arguments in order to craft more innovative and effective legal strategies for their clients.
• From a regulatory perspective, it can be used to keep on top of regulatory changes and automatically update policies and procedures to ensure compliance.
• There are already examples of lawyers adopting or creating their own legal chatbots to help free up lawyers to concentrate on more complex and creative tasks.

Broadly, “use cases” for Gen AI can be split into internal and external uses. Internal use cases being those used by lawyers “behind the scenes” to assist and improve the work product delivered to clients. External use cases include products provided by law firms to clients directly, without any further intervention by the firm. Our own research suggests that (at least for now) the use of Gen AI by lawyers is more focused on internal use cases. However, we expect that to evolve and develop as the profession moves through a period of experimentation, and more tried and tested products/solutions become widespread.

How might Gen AI change the professional indemnity claims landscape for lawyers?

For now, Gen AI has its limitations. The most obvious is that it is only as good as the data it works from. If that data contains inaccuracies or biases, the output produced by Gen AI may be compromised. In addition, Gen AI models can “hallucinate” if they cannot find an answer to a question from the data they are trained on. As a consequence, Gen AI models (particularly large language models) can produce highly convincing but inaccurate results.

In addition, Gen AI does not have “human” characteristics. It cannot exercise discretion, judgment, understand nuance, empathy or emotion. Nor can it understand a client’s specific needs or the broader context in which a piece of advice is required.

As a result, a lawyer who relies on the output produced by a Gen AI tool to provide advice to a client, without properly checking its accuracy and suitability, could be at risk of a claim.   

Lawyers also need to be conscious of their duties of confidentiality towards clients and run risk of data protection claims if a client’s data is put into a Gen AI tool without proper safeguards in place.  Similarly, there are risks to lawyers of breaching copyright if the output from publicly available larger language models (in particular) is not checked and adapted.

As things stand therefore, the greatest risk for lawyers (and their insurers) is that lawyers will adopt and use Gen AI without properly understanding how it works, the current limitations of the technology, and the need to have proper checks and balances in place to assess its output.

So, whilst the use of Gen AI may change the types of mistakes that lawyers make, the risk to them and their insurers remains (at heart) the same. At least in the short term, the claims we expect to see in the future will still arise from failures in risk management and supervision.  

That said, the use of Gen AI may well help to reduce some types of claim. There are already some tasks that Gen AI can do at least as well as a lawyer and that will only increase as the technology develops.  In particular, Gen AI tools are able to analyse vast quantities of data and documentation, reducing the risk of information being missed through a human review. Over time, the increased use of Gen AI to analyse and assist with the drafting of contracts, leases and other legal documents may also end up reducing claims in those areas.

Another common risk area for lawyers and their insurers is cyber fraud. The emergence of Gen AI poses a new challenge for email security, with fraudsters using it to create ever more sophisticated and customised phishing scams. Whilst training staff to spot something that looks suspect will still be vitally important, Gen AI can also help to combat the risk of cyber fraud. Law firms are increasingly adopting AI tools which use advanced machine learning and behavioural analysis to detect and nullify threats that traditional filters may miss.

Looking to the future

The rise of Gen AI opens up massive opportunities for members of the legal profession and their clients. Over time, it will radically change the way that lawyers work. But that change is not something to be feared. The benefits and efficiencies that Gen AI brings will help lawyers in all kinds of ways. However, Gen AI lacks human consciousness.  It cannot perform all of the roles that a human lawyer can perform. It is simply a tool which is exceptionally good at predicting and producing outputs by using the data available to it. With that in mind, it is imperative that lawyers (and their insurers) recognise its current limitations and the need for proper checks and balances to be in place around it. Having clear policies in place around how Gen AI is used, and ensuring that those using Gen AI are properly trained on its uses and limitations, will be key to ensuring the risks posed by Gen AI are mitigated effectively.

Unlocking the benefits of Gen AI but using it within a strong risk management framework should, ultimately, make lawyers a more attractive proposition for their insurers.