The rise of generative AI in law has been accompanied by both optimism and scepticism. From conferences to courtrooms, the question has been the same: can AI genuinely transform legal practice, or is it just another overhyped technology cycle?
As the founder of MiAI Law, I have seen both the promise and the pitfalls up close. The problem is not that AI has nothing to offer—far from it. The problem is that most people still assume all legal AI works the same way: a language model, trained on a large body of text, generating an answer from probabilities. This misconception is shaping perceptions, procurement and even judicial guidelines.
1. The hype cycle and why it misleads
At a recent Legal Tech Summit, I spoke with CIOs, developers and librarians who all assumed that a “legal AI research tool” simply meant sending a query to a language model. Quality, in their view, would depend on the model’s training data and the skill of the prompt. That assumption fuels both hype and distrust.
This belief persists because so many products really do operate in this way: the LLM is left to do all the work, searching, summarising and producing output. But when AI is built only on probabilities, hallucination is inevitable. A model trained to always answer will answer—even if the data is not there.
2. Doing it differently: first principles and structure
MiAI Law was designed from the ground up to do something different. We rely only on primary materials: legislation and judgments. Every case in our research base is systematically analysed and structured, identifying the ratio decidendi and doctrinal elements. That analysis becomes the foundation for how the system reasons.
When a user poses a question, MiAI Law does not let the model improvise. Instead, it follows a stepwise reasoning process, retrieving relevant cases at each stage against our own structured analysis. Only once that pool is complete is it used to produce a research report.
Those reports are not free-form summaries. They are first-principles reports, structured in IRAC (Issue, Rule, Application, Conclusion). Every proposition is footnoted to pinpoint citations, with hyperlinks to the original authority. If the material isn’t there, the system is instructed to say so.
This is not automation for its own sake. It is an attempt to mirror the way lawyers and judges reason, while preserving transparency and verifiability.
3. Why transparency matters
For courts, the key question is not whether AI can produce plausible answers—it plainly can. The question is whether those answers are reliable, transparent and verifiable. MiAI’s architecture was built to meet that test.
Like a peer-reviewed article, every statement in a MiAI report is backed by authority. The reasoning path is revealed. The distinction between ratio and obiter is made explicit. Reports are therefore audit-ready: they can be tested, challenged and trusted.
4. Credibility over hype
The distinction has not gone unnoticed. The Hon Michael Kirby AC CMG has publicly endorsed MiAI’s approach, describing it as a step-change moment for legal research. Dr Brydon Wang has recognised its commitment to trust and accountability. Senior counsel have confirmed that MiAI not only finds authorities but synthesises them— cutting research time while surfacing new arguments.
These endorsements are important not because they flatter, but because they illustrate that the method resonates with practitioners and judges who care about reasoning, not rhetoric.
5. The future: augmentation, not replacement
MiAI is not designed to write submissions, generate judgments or replace human thinking. It accelerates the first layer of legal research—the surfacing and organisation of authorities—so that lawyers and judges can focus on judgment, advocacy and strategy.
Critical thinking remains the lawyer’s task. As I often tell students and juniors: you must still read the cases, still test the reasoning, still own the analysis. AI cannot do that for you. What it can do is reduce the time spent in drudgery and create more space for creativity, strategy and justice.
6. Conclusion
Hype will come and go. What endures are the core values of our profession: accuracy, accountability and transparency. MiAI Law is not another probabilistic engine dressed up in legal branding. It is a first-principles research platform, built to be verifiable, audit-ready and aligned with judicial standards.
In an era of loud claims, that difference matters.
