Recently, I had the pleasure of speaking with Jason Marsh from AlphaLawyer. Jason and I spoke about AI and risk management. AlphaLawyer describes our conversation:
“Are you using AI in your law firm but worried it might expose you to serious risk? In this episode of the Alpha Lawyer Podcast, Jason Marsh sits down with Jennifer Ellis, attorney, legal ethics advisor, and technology consultant, to break down one of the most critical issues lawyers face today: AI hallucinations.
Jennifer explains why these errors happen, how even trusted tools can produce false information, and what lawyers must do to avoid ethical violations, sanctions, and malpractice exposure. You’ll discover why AI isn’t just another tool, it requires a completely different level of oversight. Jennifer shares real-world examples of attorneys getting into trouble from fabricated citations, explains how confidentiality risks show up in everyday use, and outlines practical strategies to safely integrate AI into your workflow without compromising your reputation. Plus, learn how to verify AI outputs properly, structure better prompts, and build an internal AI policy that protects your firm while still allowing you to benefit from the speed and efficiency of these tools.”
If you would like to listen my conversation with Jason, please check out the episode:
🎧 Apple Podcasts: https://podcasts.apple.com/us/podcast/alpha-lawyer-podcast/id1792764001?i=1000761150943
📺 YouTube: https://www.youtube.com/watch?v=GbhicTSX6OE
🌐 All podcast platforms: https://pod.link/1792764001
Here is a detailed breakdown on YouTube.
- 0:00 Introduction – The Risk of AI in Law Firms
- 2:10 What AI Hallucinations Are (and Why They Happen)
- 6:00 Why Lawyers Are Getting in Trouble with AI
- 10:15 Real Cases of Fabricated Citations and Sanctions
- 14:30 Why AI Confidence Can Mislead Attorneys
- 18:20 Verifying AI Output: What Lawyers Must Do
- 22:40 Why You Should Never Trust AI Without Checking Sources
- 26:10 The Danger of Using AI to Verify AI
- 29:30 Legal-Specific AI vs Consumer Tools (ChatGPT, Claude, etc.)
- 33:20 Confidentiality Risks and Data Privacy Concerns
- 38:00 Understanding AI Training, Memory, and Data Usage
- 42:30 How to Prompt AI for Better, Safer Results
- 46:10 Using Meta Prompting to Improve Outputs
- 49:30 Common Mistakes Lawyers Make When Using AI
- 53:00 Why Every Law Firm Needs an AI Policy
- 56:40 Malpractice Risks and Insurance Considerations
- 59:30 Training Your Team to Use AI Responsibly
- 1:02:10 Starting with Low-Risk AI Use Cases
- 1:05:00 Best Practices to Reduce Hallucinations
- 1:07:30 Final Advice: How to Use AI Without Getting in Trouble