Two years ago, when artificial intelligence platforms (hereafter referred to simply as “AI” or “AI tools”) began to be offered to professional service businesses, they were met with terrific enthusiasm, but perhaps not much seriousness. Observing these early tools’ propensity for errors (euphemistically referred to by their developers as “hallucinations”) many were lulled into a sense of security, believing that any impact on their business or career would be many years off. In reality, the transformative effects of AI have already begun to effect change.
A simple example is the increasing prevalence of AI notetakers, which replace the need for a human notetaker. This does not mean that lawyers will never again jointly appear – there are practical reasons for an associate lawyer and senior lawyer to meet a client in tandem, particularly on more complex matters. However, now that AI can act as notetaker, two outcomes can be expected:
- Without the need for this task to be completed by a person, lawyers will need to proactively discern when more than one lawyer is necessary; and
- On occasions where a single lawyer’s presence will suffice, the time previously allotted to the other(s) can be apportioned elsewhere.
Notetaking tools are a clear distillation of the types of efficiency gains available through AI which benefit both lawyer and client. Lawyers enjoy increased capabilities with an AI tool, regaining time to focus on the substantive elements of their clients’ matters, and where appropriate, clients benefit through increased cost efficiency. Notetaker tools, as a reasonably modest improvement on day-to-day efficiency, helpfully demonstrate how the change brought by AI will not take the form of conspicuous technological leaps. There are unlikely to be singular moments that inspire massive changes in how legal services are delivered. Rather, this period of transformation will be much subtler, occurring through a sustained velocity of gradual change.
That Far-Off Future Has Arrived
Notwithstanding the rapid pace at which AI tools gain capabilities, the incremental pace of this growth has disguised the potency of their potential use cases.
A startling example of this can be found in a September 2025 paper released by OpenAI in conjunction with external contributors from Microsoft, Goldman Sachs, and the U.S. Department of Justice. At the heart of this paper, titled “GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks” was a conceptually simple study: AI systems and experienced knowledge professionals were instructed to complete tasks that simulated the day-to-day work of the professionals.[1] On completion, the participants were asked to identify which result was superior, or whether the results were the same.
For the legal tasks, the work completed by AI lost by an alarmingly slim margin – with 46% of the AI-generated samples judged as being of superior or equal quality to those produced by the lawyers in the study. Further still, the tasks themselves were of considerable complexity,[2] taking the lawyer participants (who had an average of fourteen years’ practice experience) approximately eight hours.[3]
This should be sobering information for lawyers who had hoped or believed that AI could never perform some of the routine elements of legal practice. That day is here. But this should not be viewed as a threat to the profession – instead, this is a tremendous opportunity for lawyers to offer better, faster and more comprehensive services to our clients, who will continue to need the services of experienced counsel.
Lawyers themselves stand to benefit enormously, simply because the work and processes that can be augmented with AI tend to be more mundane, such as document review. Through effective AI use, lawyers will be increasing their exposure to the more intellectually challenging, interesting tasks – spending their time delivering true value to clients. During his 2017 keynote address to the UNESCO symposium on Knowledge Service and AI, Turing Award winner Raj Reddy observed that AI systems should not be viewed as replacements of professionals’ knowledge and experience.[4] Rather, AI should be regarded as a way for knowledge professionals to “amplify their cognition,” allowing them to do more of what makes them uniquely qualified.[5]
This practical approach should be embraced by lawyers; by serving clients through the amplified capabilities available through AI tools, it is possible that significant gains in service quality could be realized, while simultaneously doing so in a more financially efficient manner.
Why Hasn’t the AI Revolution Arrived in Legal Practice?
In July 2025, MIT’s Media Lab published a seminal research report entitled “The GenAI Divide: State of AI in Business 2025,”[6] analyzing the return on investment being realized by large organizations’ adoption of AI. One of the report’s findings, which quickly came to dominate all discussion following its publication, was that 95% of the enterprises which had run pilot programs using AI tools had not yet realized a positive return on investment.[7]
For AI skeptics, this finding was treated as the report’s sole conclusion. However, properly examined, the report’s conclusions are considerably more nuanced. It found that even where AI pilot programs had not yet yielded a positive return on investment, for certain task categories like document management, summaries and basic analysis, 70% of employee and managerial respondents preferred the use of AI tools.[8] The same group reported a preference for human work alone only where the tasks concerned were complex, multi-week projects.[9]
The GenAI Divide also found that professional services, including law, have already begun a transition towards a model of AI-augmented service delivery, with higher scores for existing implementation and financial opportunity than any sector outside of media/information technology.[10] In short – the report explicitly recognized that service professionals have some of the best opportunities through AI implementation, and that the pursuit of these has already begun. Another important data point for lawyers is the finding that where AI had been successfully integrated, client delivery models (the ultimate form in which clients receive the services) did not change, but instead, gains were realized by streamlining non-billable time.[11]
Of the professional services workers consulted, approximately 40% were using some form of AI regularly within their personal lives. But what should be very concerning for lawyers – whose duty of confidentiality must be strictly applied – the report found that more than 90% of enterprises consulted had employees who “reported regular use of personal AI tools for work tasks.”[12] This finding is critical in two respects; first, it demonstrates an unequivocal appetite for AI tools in the workplace. Second, it highlights a major liability risk for firms which “don’t use AI.” If proper tools and procedures are not provided to employees, the data shows that they will turn to unsecure tools, opening the door to a ticking time bomb of potential liability.
AI Training and Development: A Necessary Investment
Beyond the risks of employees’ use of unsecured AI tools in the absence of resources appropriate for legal work, lawyers who choose not to adapt to the realities of the AI-augmented legal marketplace are choosing to swim against an increasingly swift tide. Doing so may not be in the interests of their practice, and more importantly, in those of their clients.
In Justice Myers’ 2021 decision in Worsoff v MTCC 1168 et al, lawyers’ duty of technological competency was discussed in the context of remote hearings.[13] The decision remains noteworthy for its confirmation that the duty of technological competency is a moving target as technology evolves.[14] In 2026, we can reasonably expect that the prevalence and acuity of AI will continue to rapidly expand, and with it, the standard of reasonable technological competency which lawyers must meet. For those who delay their learning, the frenetic pace of change will make this more challenging.
The path to overcoming the attendant challenges in this period of change, as confirmed by MIT’s researchers in The GenAI Divide,[15] is through continued education and training. Hull & Hull LLP has, now for more than a year, made a concerted effort to ensure that our team has access to purpose-built AI tools for lawyers, and critically – to develop training and education to ensure that they are used effectively and safely. Our firm is not alone in this regard – some Canadian firms have announced the establishment of AI committees or executive roles.[16] In the United States, Ropes & Gray has formally begun an AI learning initiative, “TrAIlblazers,” allotting junior lawyers 400 hours of billable credit for 2026 to train on AI tools.[17]
Formalizing Training and AI Adoption: The CNSL.ai Advantage
Hull & Hull LLP’s path to proficiency was not without its challenges – change is not easy in any organization. But by concentrating on skills and resources grounded in how lawyers actually use AI, Hull & Hull LLP has built out a framework for practical innovation, which has begun to bear fruit. One visible outgrowth of this overall AI strategy can be found in some of Hull & Hull LLP’s publishing, such as the AI-powered podcast, The Hull & Hull Debrief.[18]
These concentrated efforts led to the establishment of CNSL.ai (pronounced “Counsel AI”), a company which provides AI training specifically for lawyers, based upon the experimentation, training, and development that has formed the backbone of Hull & Hull LLP’s continued AI strategy.
CNSL.ai exists to teach lawyers how to effectively and safely get the most out of AI tools within their practice.[19] Through in-person training sessions that are designed and run by lawyers, CNSL.ai has a unique advantage in understanding what lawyers and their firms need for successful adoption of AI within their practice.
The opportunities for lawyers using AI are potentially enormous – and self-reported market data from the 2025 Future of Professionals Report shows that firms supported by a deliberate implementation structure are 3.5 times more likely to realize a positive return on their investment in AI technology.[20] This is where CNSL.ai presents a significant advantage for its client firms: through a structured model of implementation and training for lawyers, CNSL.ai can sharply reduce the learning curve, allowing firms to realize the gains from AI faster, and more comprehensively. For established firms, AI adoption can be used as a way of preserving a market edge, and for ambitious, rising legal teams – adopting these systems can be the fulcrum with which larger incumbents can be moved.
This is an exciting time to be a lawyer, and undoubtedly, 2026 will hold a number of fascinating changes. If you or your firm are interested in having a consultation with CNSL.ai to see how AI tools can help your firm, please reach out to the CNSL.ai team at info@CNSL.ai, or by telephone at (647)-256-1218. Tremendous opportunities await those willing to seize them – for firms ready to move from pilot programs to measurable results, CNSL.ai can be your guide on this path.
Ian Hull, LSM
Partner, Hull & Hull LLP
CEO of CNSL.ai
Editor: Suzana Popovic-Montag
spopovic@hullandhull.com
141 Adelaide St. W. Suite 1700, Toronto, Ontario
M5H 3L5 Tel: 416-369-1140
[1] Tejal Patwardhan et al, “Gdpval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks” (2025) [preprint, OpenAI and Cornell University] at 4.
[2] Ibid, 27.
[3] Ibid, 2.
[4] Guardian Angels and Cognition Amplifiers: AI 2.0 Technologies in Support of Humankind, Keynote Speech at UNESCO Symposium on Knowledge Service and AI (Hangzhou, China, 28 September 2017).
[5] Ibid.
[6] Aditya Challapally, Chris Pease, Ramesh Raskar & Pradyumna Chari, The GenAI Divide: State of AI in Business 2025 (MIT Media Lab, July 2025) at 3.
[7] Ibid.
[8] Ibid, 13.
[9] Ibid.
[10] Ibid, 5.
[11] Ibid, 6.
[12] Ibid, 8.
[13] Worsoff v MTCC 1168, 2021 ONSC 6493 (CanLII), online: https://canlii.ca/t/jjf53 at para 32(g).
[14] Ibid.
[15] Challapally et al, The GenAI Divide, supra note [6] at 20.
[16] Frédéric Wilson & Simon du Perron, Decoding Tomorrow: BLG Primer on AI Governance (Borden Ladner Gervais LLP, 19 December 2023), online: https://www.blg.com/en/insights/2023/12/decoding-tomorrow-blg-primer-on-ai-governance.
[17] Sara Merken & Mike Scarcella, “Law Firm’s AI Experiment Gives Lawyers a Break from Billable Hours”, Reuters (6 November 2025), online: https://www.reuters.com/legal/government/law-firms-ai-experiment-gives-lawyers-break-billable-hours-2025-11-06.
[18] Hull & Hull Debrief Series (Hull & Hull LLP), online: https://hullandhull.com/podcasts-audio.
[19] CNSL.ai Inc, “Our Methodology” (20 October 2025), online: https://cnsl.ai/methodology.
[20] The Future of Legal Professionals Report (Thomson Reuters Institute, 26 June 2025), at 10.
