Lawyers' biggest barriers to AI adoption

Generative AI adoption accelerates

The appetite for generative AI technology in the legal sector is unprecedented, with lawyers from all backgrounds jumping at the chance to make the most of its time-saving potential.

In January 2024, LexisNexis surveyed more than 1,200 legal professionals in the UK on their use of generative AI. A little over a quarter (26%) of respondents revealed they're using generative AI tools at least once a month. This is a noticeable rise from only 11% of respondents in July 2023.

A quarter (26%) of legal professionals are using generative AI at least once a month.

Respondents from academic institutions and large law firms were the most likely to use these tools for their work, at 33% and 32% respectively.

Generative AI will fundamentally change the way we work in the legal industry, says Dr. Gerrit Beckhaus, Freshfields Partner and Co-head of the Freshfields Lab.

"Generative AI has the potential to support a lawyer's work by helping with efficiencies, increasing effectiveness and speed, which will lead to a superior output for the client," he says.

While most lawyers are eager for the time-saving – and, potentially, cost-saving – benefits that come alongside generative AI, concerns still linger. And rightly so...

Excitement levels may be high, yet many lawyers still have concerns over the ethical implications of AI. The technology has a tendency to fabricate answers, leak confidential data, and construct biases, posing an obvious challenge for the legal profession.

Only 10% of respondents to our survey said they had no concerns using generative AI. Most respondents (59%) admitted they had some concerns, while just under a third revealed significant (26%) or fundamental concerns (6%). This is almost identical to findings from the July 2023 LexisNexis generative AI survey.

Mark Smith, Director of Strategic Markets at LexisNexis, says fears around accuracy of information are – and always will be – of vital importance to the legal profession.

"There's a huge fear of providing the wrong advice to clients, which is particularly heightened by issues with free-to-use generative AI."

The biggest hurdles to adoption of generative AI are content hallucinations (57%), security concerns (55%), and the technology not being trustworthy enough (55%).

How to manage the risks of artificial intelligence in your business

The biggest hurdle to generative AI adoption for lawyers is content hallucinations.

According to Joe Cohen, Director of Innovation at Charles Russell Speechlys, lawyers can avoid hallucinations on free-to-use AI tools by double-checking all outputs and verifying them from other sources.

"There are also certain use cases that are not recommended on free-to-access tools. For example, asking AI to cite cases for you."

Beckhaus from Freshfields says hallucinations are intrinsic to generative Large Language Models (LLMs), which are trained to be creative and avoid repetition.

"There are technical ways to reduce the tendency of hallucinations by increasing the parameters to make generative AI more deterministic, such as lowering the temperature."

The "temperature" is a hyperparameter that can be used to control the randomness and creativity of generated text in generative language models.

"In addition, prompting and context can further reduce the risk, alongside human guidance that can help to validate, assess and contextualise the results produced with the help of generative AI."

Find a legal framework and regulatory guidance around explainability on Lexis+.

LexisNexis' Smith says there are deeper, more structural challenges organisations need to overcome before adopting and deploying AI solutions.

"Firms need to consider everything from their pricing and resourcing models, to the investment costs for acquiring, developing or training generative AI solutions, to employee concerns around job security and satisfaction."

Besides inaccurate or unreliable information, another major risk of external firms using free-to-use generative AI tools is potential data breaches, says Bunmi Tandoh, an in-house solicitor at London Borough of Enfield.

"Possible data protection breaches and breaches of confidentiality are a major risk," she says.

See guidance notes on data protection for artificial intelligence on Lexis+.

May Winfield, Global Director of Commercial, Legal and Digital Risks at internationally renowned engineering advisory firm, Buro Happold, says there is also the issue of confidentiality. 

"Inputting confidential data onto a free-to-use generative AI platform is like throwing it into a public forum; you can’t delete it or remove it and someone asking the right question could, potentially, extract that data." 

While these risks are worrisome, some would argue that perhaps the biggest risk to lawyers is failing to make use of generative AI.

Head of Lawtech and Chief Knowledge & Innovation Officer at Macfarlanes, Chris Tart-Roberts, says, "AI is shaping the legal industry and no business or industry can afford to ignore it given its transformative potential."