AI: friend or foe? By Steve Brett, co-founder of E3 Compliance Training and executive board member for the UKLTA (UK LegalTech Association).
A poll of 700 companies by the British Chambers of Commerce in September found that just under half had no plans to use AI, a figure that rose to 58 per cent for businesses selling to consumers. This is despite reports that in August 180.5 million people worldwide were using ChatGPT, one of the leading generative AI chatbots that have emerged this year.
I suspect many conveyancing firms fall into the above category. Unsure of how AI fits into their legal world and rightly concerned about the myriad of associated privacy and data complexities that it brings. With large law firms experiencing on average two ‘cyber incidents’ a month, conveyancers are right to be cautious.
However, these technologies are here and they are developing within the commercial world at an increasing pace. For conveyancers, it’s wise to consider now how this technology will evolve into the workplace and what limits may be needed to ensure client and firm data remains secure.
Today it’s important to understand what the AI tools are designed to do. They may appear like magic but like any tool it is important to understand the power and limitations. Most conversation today is about ‘large language models’ which are often referred to as LLMs. Examples of this type of tool include CHAT GPT, Bard and Bing AI. All of these are designed to produce plausible sounding responses to a question and are most frequently seen deployed on consumer facing websites.
Anyone who says that artificial intelligence tools will not be able to do a task are almost certainly wrong – it’s just a matter of time. So how should conveyancing firms look at the adoption of AI tools today?
The first step should be to ensure that AI tools are not inadvertently abused. There are numerous examples where an employee has used an AI tool to solve a problem and their well-intentioned actions have gone horribly wrong. Do your team know not to include confidential information into a public AI tool such as ChatGPT? An employee at Samsung did not anda by asking a question about a piece of code ended up publicly disclosing sensitive intellectual property. There are also examples of employees who have entered personal data resulting in GDPR breches.
If you have not done so, the initial step should be to ensure that all staff know they should not be using any AI based tool for work purposes that has not been agreed by the firm. Managers need to explain that these tools have issues that can catch out the unwary, and must be used effectively. There are already several cases where a lawyer has used an AI tool to prepare court documents that included totally fictitious case references which, at the least, are embarrassing for both the individual and the firm.
We will see many tools become ‘AI enabled’ over the coming months and years. Some of this will be the real application of AI techniques but others will be better thought of as the abuse of marketing speak. Be alert to these changes and make sure that staff know that they can either use these features or have them disabled.
AI tools are going to be part of the future, so you need to ensure that these tools are adopted consciously rather than by accident. The key will be to develop some organisational knowledge and expertise. This could be via an employee with the technical interest who is given the training and time to explore the tools today. There is nothing better than trying the tools firsthand to safely explore how they will work in the ‘real world’. This experience will highlight how the tools can benefit your firm and where they are not yet ready for day to day use.
Spend time with your existing suppliers and potential new suppliers to understand how they are using technology to solve your problems. For example, what training data has been used, what biases were expected and how were these minimised? How will the training be managed to stop the model producing incorrect or unwelcome responses?
One key thing to remember is that many applications will require the development of new skills. For example, the responses to questions can be adapted by careful use of so called ‘prompts’ which is becoming a sought-after skill giving rise to the new job of ‘prompt engineer’.
AI tools are going to have a big impact on how we work in the future. Take careful, thoughtful steps and you can ride the wave of change while integrating elements of AI that work best for your business.