According to Cesar Taurion, a former IBM executive and today one of the greatest evangelisers of Artificial Intelligence in Brazil, the path to the proper use of AI tools must follow certain key points. We have adapted some of these points for law firms, including:

1 – Incorporate AI into DNA

Incorporate AI (AI in general, not just generative AI) into the business strategy. The office shouldn’t have an AI project to be cool or to demonstrate to the market that it’s on top of the trends. Its main objective should be to solve a real business problem, and that problem can be solved with AI if it really is the best solution.

It’s an example of adopting an AI tool to make M&A due diligence more effective, or jurimetrics and data analysis to make an audit report more assertive.

2 – Objectivity

Have clear objectives for AI projects. Without ROI, projects will only live on hype, and hype is short-lived and leads to frustration, says Taurion.

Take the case of hiring the most robust and expensive AI tool on the market: if you don’t have an action plan and a defined objective, it will be a wasted investment.

3 – Evangelise members

Without leadership interaction and support, AI projects will be mere operational projects, losing much of their transformative potential.

In other words, you have to win over the Board of Directors and, above all, the initial barrier of possible scepticism. As Isabel Parker, winner in the professional services category of the European Women of Legal Technology Awards 2020, rightly points out, law firms in general are making money, which suggests that the business model works. So why tinker with something that isn’t wrong? Incidentally, the term “reinventing” has been adopted by global law firm Baker & McKenzie to name its innovation projects.

4 – Find and retain talent

We need to be careful when we say that ChatGPT and the like won’t require as many talents, including those with in-depth knowledge of AI. On the contrary: they are still needed! That’s why selection processes, especially for “fertilised and creative minds” and young talent, should have sex appeal and win over this audience. So traditional recruitment is out of the picture and processes that emphasise solving contemporary problems and challenges are in favour.

5 – Implement data governance

The old maxim “rubbish in, rubbish out” is still in vogue, and according to Taurion, data has to be structured. That’s why it’s essential to have a knowledge management team in line with the best market practices. A database, especially of contracts, will help a lot with indexing and the much-vaunted language model training.

6 – Invest in technological infrastructure

Computing capacity, whether in-house or external, in the cloud is indispensable and has to be constantly adjusted. Investment is needed, because after all, nothing is free.

7 – Create an AI governance committee

Don’t underestimate issues of privacy, ethics, regulation and copyright infringement, especially the use of sensitive data when training language models.

Any failure can cost a lot, not only in money, but also in image and reputation, especially coming from law firms.

Attention should also be paid to compliance and regulatory aspects, particularly those of the Bar Association. In several countries, the respective class sessions have already issued standards and rules of conduct regarding the adoption of AI in general and not just generative AI.

In addition, clients generally operate in different countries with different laws. Here, then, is the old maxim that lawyers should set an example at home.

8 – Monitor results and mitigate biases

As soon as a machine learning system goes into production, it starts to degrade. Keep the algorithms adjusted and validate your results to mitigate biases and possible “hallucinations”, typical of systems like ChatGPT.

Imagine the stain this could have on the image of lawyers. It’s worth remembering the US episode where a lawyer cited a non-existent court precedent after extracting the information from ChatGPT.

9 – Technology is not at the centre

It is essential to review processes, train people and clearly explain the objectives and benefits of the solution for the company, its clients and employees.

Therefore, workshops, simulations and debates, such as the adoption of the best prompts and how certain tools can automate certain repetitive tasks, the preparation of audit reports and chatbots for trivial customer questions, are some of these examples.

10 – Be patient

The value generated does not appear immediately. It’s a long-term project, full of intangible issues and often without visible metrics. In any case, the arrow of time is forward. It is inevitable that law firms will embark on a path that adopts AI.

After all, as José Saramago used to say, let’s not rush, but let’s not waste time.