The world has been experiencing the breakneck pace of the 4th Industrial Revolution, marked by emerging technologies such as artificial intelligence (AI), neurotechnology, automation, blockchain technology, cryptocurrency, and others. It is also said that we are already seeing some of the signs of the 5th Industrial Revolution with advancements in fields like AI and robotics, and with company adoption of automation and machine learning to enhance productivity and efficiency.
With that, very soon, there may not be many tasks or functions for which there isn’t an AI tool. And this becomes apparent if one looks at the website There's An AI For That® — Find The Right AI Tool For Any Task. So, despite any angst, anxiety or fear some of us in the legal profession may have, AI seems to be here to stay. As responsible legal practitioners with a professional code, attorney-client privilege and data privacy laws to uphold, if AI use is permitted in our jurisdiction or by our Bar Association, and if we choose to use it for our work, we have a responsibility to understand the technology to be able to use it responsibly.
Similarly, law students also need to abide by any policy which their university or law school has established in relation to the use of AI during their course of study, and their understanding and familiarity with AI will likely aid them in their future work.
A bit about AI
AI involves the use of computers or machines to perform tasks or make decisions that are usually associated with human intelligence. Natural language processing allows computers to transcribe normal human translations but to also gain context and to some extent understand it. This occurs in Alexa and Google Assistant, for example.
In AI, pattern recognition is the ability to identify regularities and relationships in data. It allows AI systems to analyze data, classify information, make predictions, and ultimately, make decisions. Pattern recognition in AI focuses on finding recurring structures, trends, and relationships within data, whether it's visual, auditory, textual or numerical. This is achieved through the use of algorithms that are trained to detect specific features or patterns within datasets. Pattern recognition is used in AI for uses such as natural language processing, image and facial recognition, fraud detection in financial transactions, medical diagnoses, and autonomous vehicles.
AI in the legal profession
Some of the benefits of using AI in legal practice include the automation of routine tasks, efficiency, cost reduction, the processing of vast amounts of data and the extraction of insights from said data. Using AI in legal practice comes with some risk, such as hallucinations (production of fabricated or inaccurate information by an AI system), bias (as a result of the exclusion of people or a subset of people from the AI model), data privacy risks (as a result of the transfer of confidential client data out of our legal practices/organizations) and the explainability problem (there’s no explanation for how an AI platform came up with an answer).
Sources of information for AI
To better understand and use AI responsibly, understanding some of the sources of the information is important.
ChatGPT is said to be trained on information from the internet up to 2021 and various research publications. Additionally, it was reported that for Google AI Overviews, the most cited domains around June 2025 were: quora.com, reddit.com, linkedin.com, youtube.com, nytimes.com, amazon.com, travelandleisure.com, healthline.com, bankrate.com, forbes.com, nerdwallet.com, goodhousekeeping.com, runrepeat.com, cnet.com, businessinsider.com, medium.com, en.wikipedia.org, foodandwine.com, pcmag.com and yahoo.com.
This may be a fair indication that if a lawyer is solely relying on a general-purpose AI platform for legal research, without verification and fact-checking, the likelihood of preparing legal work products that contain errors may be quite high.
Smart & Responsible use of AI
Here are some general, non-exhaustive measures that legal practitioners can take to use AI responsibly:
Read and understand the terms of use, data privacy and data retention policies of the AI platform. Deselect or take off any data retention feature in the AI platform.
Use platforms that are more sophisticated (such as Perplexity) or industry-specific (such as Lexis AI or vLex), and learn to use them well. In the paid version of Perplexity (Perplexity Pro), you can select your sources, e.g. financial (SEC filings, company financial disclosures), academic (research papers, scientific studies), social (community recommendations, reviews, trending discussions) or the entire web (real-time news, current events).
After using a general-purpose AI platform, you may then resort to looking for the material or cases cited in the AI output to further your research and reading.
Your output is only as good as your input. AI does not (currently) understand the world or broad implications, just the information or questions you bring to it. So, pay attention to the inputs you use to obtain your outputs. Be mindful that AI outputs can contain the verbatim text of third-party sources, and this could have copyright and other legal implications. Citation checking is important to give proper attribution to third-party sources.
Verify and fact-check your outputs. Look for citations in-line so you can determine if the AI platform is hallucinating or not. You can also understand the language model’s thought process.
Upskill in areas like prompt engineering so you can better help to guide the AI platforms that you use. Avail yourselves of the various free or low-cost courses or webinars that are offered to learn about and understand data, Big Data, language models, AI and related areas better. Some of these courses are offered by IBM, Microsoft, EdX, and Open AI.
No more AI hallucinations
If we put some of the above into practice, our profession will hopefully see fewer incidents of hallucinations arising in court cases. Recent examples of this can be found in cases in Trinidad & Tobago CV2023 – 04039 Nexgen Pathology Services Ltd v Darceuil Duncan, the Bahamas Marcus Gruesser & Ira Hess v Robert Phelps Herman 2023/PRO/cpr/00006, and in the cases of Mata v. Avianca, Inc. No. 22-cv-1461 (PKC) (S.D.N.Y.), United States v. Cohen, No. 1:18-cr-00602-JMF (S.D.N.Y. Dec. 12, 2023) and Zhang v Chen, 2024 BCSC 285.
Going forward
AI is not just a computer science, engineering or technology thing, it is an all-disciplines and all-professions thing! With AI becoming more commonplace in our profession, we need to equip ourselves so that we can use this technology in a smarter, more informed manner. Where AI is being used, it must be trusted and verified. While we may be curious to use the technology, don’t lose sight of ethics, accuracy, reliability and governance. Be a reflective and responsible lawyer by improving your knowledge about this and other evolving technologies. In this way, we can uphold our professional code, our duties to our clients, abide by the law, and assist our courts in upholding high standards in the administration of justice as the world proceeds from the Fourth Industrial Revolution and into the Fifth.
Bellina Barrow is the Principal Attorney/Founder of Tenoreque Legal, a virtual legal practice based in Trinidad & Tobago since 2021. A former legal tutor and a dedicated mentor, Bellina is committed to fostering and contributing to thought leadership in law, fintech, tech, digital assets and sports by deconstructing and demystifying these areas via practical and digestible storytelling and writing. Outside of her technical and academic writing, Bellina is also a co-author of the books Soul of An Athlete (2023) and Women in Law: Discovering the True Meaning of Success (2022).
