The Role of AI in Disputes

Rapid developments in GenAI continue to transform the dispute resolution landscape, as lawyers seek to leverage these tools for a wide range of uses throughout the lifecycle of a claim.

GenAI's impact to date

Advances in GenAI and the large language models (LLMs) that power them have led to a proliferation of specialist tools to assist parties with various aspects of dispute resolution.

Lawyers are able to use these tools to intuitively carry out document-related tasks such as summarising documents, answering questions on a document set, extracting key information, identifying inconsistencies, and searching for and categorising documents for relevance or by issue – all in a fraction of the time it would take for a human.  LLMs are also able to assist with drafting any kind of text with suitable prompting – including correspondence, chronologies, statements and submissions. LLMs are increasingly being used to streamline legal research, with specific products linking directly to the latest cases and legal developments.

LLMs are not limited to language – products that work with sound and are able to provide summaries and transcripts of calls are already in use. LLMs can equally work with images, both still and live footage, if needed.

The adoption of these tools is leading to significant efficiencies, particularly for larger-scale, evidence-heavy disputes. To date, these tools have had the greatest impact in often one of the most time-intensive and costly aspects of a dispute – disclosure. LLMs are being used to dramatically cut-down on the time required to search for and identify relevant documents, as well as assist in the more nuanced legal tasks of redacting privileged or confidential and irrelevant information. GenAI tools also have a wide range of applications when it comes to analysing large volumes of documentary evidence such as summarisation, trend or gap analysis, timelines and so on. Such functionality is particularly useful for parties dealing with large volumes of previously unseen disclosure from the other side in a dispute.

With great power, comes great responsibility

As with any new technology, GenAI's adoption within disputes has not been without incident or risk. Many will have seen the headlines of unchecked 'hallucinations' created by GenAI resulting in incorrect material, including fabricated legal authorities, being put before a court.

The reality remains that even market-leading GenAI tools can, on occasion, generate incorrect information (hallucinate). This is an inherent risk with the LLMs that are currently available. Whilst extremely advanced, the output from these tools is ultimately the product of (in essence) probability, rather than any conscious thought or analysis.  This is far from a fatal flaw. For certain tasks, AI can still outperform human beings when it comes to margin for error.

Courts have consistently emphasised that responsibility for the accuracy of the information put before them rests with the lawyers producing it.  Firms are often addressing these risks through specific AI use policies and training, requirements to verify AI-generated underlying analysis against primary source material, and instructions and guardrails requiring the LLM to cite sources for every answer it gives.

In terms of security and data protection, special enterprise versions of LLMs promise a secure environment for client data where nothing is used to train the underlying LLM. The use of such enterprise tools has become all the more important in circumstances where some courts have suggested that the use of non-confidential, public AI tools could undermine a party's claim to privilege over any information uploaded to them.

Looking to the horizon

GenAI remains an extremely fast-developing area. In particular, with the advent of agentic AI, it is likely that tools will be developed to deal with increasingly complex aspects of dispute resolution and to do so in an autonomous or largely autonomous manner. This appears likely to extend beyond the now more established areas of documents analysis and drafting, to the more qualitative tasks of assisting with merits assessment and the formulation of legal arguments.

The English courts continue to encourage the use of technology, including GenAI, to reduce the costs of litigation, whilst consulting on what further rules (if any) should be introduced to manage the associated risks.