AI False Citations “Hallucinations,” and Legal Risk: What Every Lawyer Needs to Know About Verification and Reliability

Ali Mohammed | JuristAI Legal Technology Group, Inc

Ali Mohammed is the CEO and a co-founder of JuristAI. He has always had a strong sense of justice and a passion for the law. Ali volunteered as an intern at a criminal justice firm during college. He went on to achieve a master’s in IT Management and became a solution architect.

Mark Underwood | JuristAI Legal Technology Group, Inc

Mark Underwood is the Co-Founder of JuristAI and creator of Lawvocate, an AI legal agent he trained that runs as a private GPT on OpenAI GPT-5 Thinking and is rolling out via the OpenAI API. A Fractional Chief Legal Officer and the Legal Community Leader at Startups.com, Mark advises founders and legal teams on AI governance, data, IP, and contract strategy. He has 37+ years of experience aligning product velocity with legal defensibility and is admitted to practice before the U.S. District Court for the Western District of Michigan.

Mark’s practical frameworks and commentary on law and emerging tech routinely reach thousands of practitioners and operators across the startup ecosystem.

Agnes Malatinszky | JuristAI Legal Technology Group, Inc

Agnes Malatinszky is a seasoned operator, investor, and leader who now builds and advises companies at the intersection of law, technology, and human centered design. A licensed attorney with a passion for practical problemsolving, she has helped scale organizations to millions of users and now focuses on creating entrepreneurial pathways and AI-powered tools that help people thrive.

I. What is an AI hallucination? | 1:00pm – 1:10pm

II. Why do hallucinations occur? | 1:10pm – 1:20pm

III. When are users most at risk of hallucinations occurring? | 1:20pm – 1:30pm

IV. What kind of AI produces hallucinations? | 1:30pm – 1:40pm

V. The categories of hallucinations (fake cites vs. pincites vs. fake quotes vs. irrelevance) | 1:40pm – 1:50pm

VI. Consequences of submitting AI hallucinations in court (Rule 11) | 1:50pm – 1:55pm

VII. The AI Hallucination Cases Database (https://www.damiencharlotin.com/hallucinations/) | 1:55pm – 2:00pm

Break | 2:00pm – 2:10pm

VIII. When AI can be trusted and when it cannot | 2:10pm – 2:15pm

IX. Is this problem unfixable? | 2:15pm – 2:20pm

X. Is it worth using AI if this is a risk? | 2:20pm – 2:25pm

XI. Automated post-processing of AI output | 2:25pm – 2:30pm

XII. Manual post-processing of AI outputs | 2:30pm – 2:40pm

XIII. How JuristAI pre-empts and removes hallucinations from our work outputs | 2:40pm – 2:50pm

XIV. How lawyers can build a manual verification workflow (source of-truth retrieval, citation validation, and human-in-the-loop review) if they aren’t using software hallucination detection | 2:50pm – 3:10pm