Recent coverage of AI-generated hallucinations finding their way into court filings is a signal the legal industry shouldn't ignore. These aren't just edge cases; they're a preview of what happens when AI is deployed without the right structure and data sources around it.
We want to address this directly because we think about it, and the implications for our customers and their clients, every day.
Legal is an extremely high-stakes environment for AI. Case records are often complex, and legal filings contain a lot of nuance. That reality has shaped every product decision we've made at Supio.
What Hallucinations Are and Why They Matter in Law
A hallucination, simply put, is when AI produces something false as if it were true. Not because it's trying to deceive anyone, but because AI models are built to produce fluent, confident-sounding responses, even when they're filling in gaps they don't have the answer to.
In many contexts, those hallucinations are inconveniences, but in the legal world, where facts and accuracy are everything, they have the potential to destroy a case, ruin a firm’s reputation, or cost an attorney their license.
Here’s the truth about AI: whenever you feed something into a model, there is always a percentage chance that the output is wrong. The question isn’t whether that risk exists. It’s whether you’ve built the right checks around it. Client trust hinges on an attorney’s ability to verify what their tools produce, and that starts with choosing an AI partner that builds verification and trust directly into workflows.
For most outputs in a case, that means source-linking every fact to the document it came from so attorneys can verify what they are looking at. For legal citations and case law, source-linking alone isn’t enough. AI can confidently cite a case that doesn’t exist, or misstate a holding in a way that looks credible on the page. The only real check against that is access to an authoritative source of truth for the law itself.
That’s the problem our partnership with Thomson Reuters is built to solve.
What the Supio and Thomson Reuters Partnership Means in Practice
Earlier this month, we announced the next step in our partnership with Thomson Reuters, providing a direct channel for Supio users to access Westlaw Advantage without leaving the platform.
Westlaw Advantage is the definitive authority for legal research and case law. When an attorney is drafting inside Supio and needs to verify a citation, that verification no longer requires switching tools, reframing the research, or remembering to check. One click, and that authoritative source is built into the workflow.
The integration connects Supio and Westlaw Advantage, bridging case building capabilities with legal research verification data at three specific moments in case development: when jurisdiction-specific research would strengthen the analysis, when an attorney needs to go deeper on a legal issue, and when a draft needs to be validated before it moves forward. This helps eliminate context switching, which is often where errors enter a workflow.
When attorneys move between disconnected tools, gaps open up. Case facts and strategy don’t connect cleanly to legal research, and the risk of error compounds.
Supio has always been built on the idea that AI outputs must be auditable and verifiable, not trusted blindly. The Westlaw Advantage integration extends that principle to the legal research layer. It's the next step in a partnership we're committed to building on.
AI Doesn't Change an Attorney's Ethical Duty
Attorneys have always had to verify the accuracy and veracity of their filings, whether it was a first-year associate or a brand-new paralegal who prepared the work. That obligation doesn’t change with the introduction of AI.
AI is becoming an essential tool in case building, but it isn't a replacement for the human aspects of legal advocacy—the judgment calls, the strategic decisions, the responsibility to the client. We have to be honest about what AI can do. It cannot take fiduciary responsibility off an attorney’s plate. It never could, and it never should.
This is the reality that we’ve built Supio around. Every output is grounded in case facts and traceable to the source. And now, with Westlaw Advantage accessible directly from the platform, attorneys have an authoritative voice in legal research available at every step of the workflow.
Verification isn't a step you add after using Supio. It's built into how Supio works.
What We’re Building Towards
The legal industry is at an inflection point. Courts are grappling with AI in filings, and attorneys are actively being sanctioned for submitting AI-generated content they failed to verify. The bar for responsible AI use is rising, and it should.
This is why an open, integrated system matters. Attorneys building cases in completely closed platforms increase the risk of unverified data making its way into filings. Connection to authoritative sources isn’t a nice-to-have, it’s table stakes.
We believe the firms that will lead are not the ones that adopt AI fastest. They're the ones who adopt it most carefully, with systems that make verification easy, outputs that are auditable, and a clear understanding of where human judgment remains essential. Its power is in making something good, better or more efficient, but its output still requires checks and balances by those using it.
That's what we're building. We’re doubling down on solutions that deliver real value and AI automation where it counts most—all while reducing risk or second-guessing that’s often associated with AI. And we're committed to being transparent about both the promise and the responsibility that come with it.
— Jerry Zhou, Co-Founder & CEO