In just a few years, generative AI has moved from being a novelty to a tool that is reshaping the way many of us practise law. What began in late 2022 as an experiment is now part of daily professional life. Whether it is producing a first draft more quickly, refining an affidavit, or simplifying client communications, AI has become a practical support for the modern lawyer.
Recent research puts this trend into perspective. By mid 2025, ChatGPT was being used weekly by nearly 700 million people, “around 10% of the world’s adult population,” and handling billions of messages every day.
For lawyers, the most relevant finding is how central writing with ChatGPT has become. The study reports that “writing is the most common use case at work, accounting for 40% of work-related messages”. And, importantly, most of those uses are not about drafting from scratch. Instead, they involve “editing, critiquing, [and] translating text provided by the user” This mirrors the reality of practice, where lawyers often rely on AI to polish factums, refine affidavits, or turn rough notes into something that can be sent to a client or filed with the court.
Another key finding is ChatGPT’s role in decision support. Nearly half of all professional use falls into the category of “asking”; seeking explanations, background, or guidance to support judgment, For lawyers, this might mean testing procedural checklists, reviewing statutory language, or brainstorming strategy in a complex matter. The value lies in improving decision making, not in replacing it.
But there are risks when technology is not properly supervised. Ontario’s recent decision in Ko v. Li illustrates this clearly. As my colleague Shawnee Matinnia explained in her blog, Justice Myers required counsel to show cause after discovering that a factum filed with the court contained nonexistent authorities, later admitted having been generated with AI assistance. Although the lawyer apologized, committed to stronger verification processes, and the contempt order was withdrawn, the case highlights the dangers of relying on AI outputs without review. The Rules of Civil Procedure are clear that counsel must certify the authenticity of every authority cited in a factum. AI does not change that responsibility.
We are still at an early stage of integrating AI into legal practice. Ko v. Li shows that courts are already beginning to address its risks, but this is only the start. As lawyers, judges, and regulators respond to the growing use of AI in daily work, more case law, professional guidance, and ethical standards will continue to develop. What is clear is that ChatGPT and similar tools are no longer a passing trend. They are becoming part of the way society and the courts are being reshaped. The challenge is to make sure this development strengthens rather than weakens the integrity of the justice system.
Thanks for reading!
Grey Tiansheng Wen
References:
Aaron Chatterji, Tom Cunningham, David Deming, Zoë Hitzig, Christopher Ong, Carl Shan, and Kevin Wadman, How People Use ChatGPT (15 September 2025), at 1.
Ibid, at 2–3.
Ibid, at 16–17.
Ibid, at 17.