Earlier this week, I covered The Coming Use and Misuse of Artificial Intelligence in the Courtroom. Here are five takeaways from the article.

The article by Florida Chief U.S. Magistrate Judge William Matthewman is one of six articles in a special E-Discovery issue published by the Journal of Technology Law & Policy in partnership with the University of Florida’s E-Discovery Conference. I provided a summary of the article (available here) earlier this week. Here are five takeaways from The Coming Use and Misuse of Artificial Intelligence in the Courtroom.

Takeaway 1: The “Liar’s Dividend” and the Rise of the Deepfake Defense

As public awareness of generative AI grows, a parasitic litigation tactic has emerged: the “Deepfake Defense.” In this scenario, a party does not need to prove the evidence is fake; they only need to create enough doubt to make the truth unusable.

By baselessly asserting that legitimate, incriminating evidence is a deepfake, litigants exploit the “Liar’s Dividend.” Judge Matthewman contends that this is a dangerous erosion of the moral fabric of the law, where the mere existence of AI technology provides a shield for the guilty by delegitimizing the authentic.

Takeaway 2: The Malpractice of Avoidance

While the misuse of AI is a primary concern, the refusal to engage with it presents its own ethical crisis. Judge Matthewman highlights a counter-intuitive evolution of the law: failing to use AI may soon constitute legal malpractice.

Referencing Judge Jesse Furman’s perspective, the article contends that billing thousands of hours for tasks a machine could complete in thirty seconds is increasingly unreasonable. But this goes beyond billing. Technical proficiency has become a core component of the ethical duty of competence. Avoiding AI isn’t just inefficient; it’s a failure to provide competent representation in an era where technology is the benchmark for reasonableness. But AI remains a double-edged sword, offering pro se litigants a path to justice through clearer arguments while simultaneously threatening the court with “AI-fueled hallucinations.”

Takeaway 3: Unacknowledged Evidence is the Tip of the Iceberg

Judge Matthewman makes an important distinction between acknowledged AI (evidence disclosed as AI-enhanced) and unacknowledged AI (fabricated evidence).

The hallucinated citations currently plaguing legal briefs are merely the tip of the iceberg. The true threat lies in the unacknowledged fakes that evade scrutiny entirely. The Western District of Texas recently saw an innocent person indicted based on fake audiovisual evidence provided by a confidential informant. The fraud was only discovered because the informant happened to plead guilty in a completely separate matter. This highlights the challenge that “we don’t know what we don’t know”: it’s likely that individuals are currently being prosecuted, or even convicted, based on unacknowledged fakes that have successfully bypassed current judicial filters.

Takeaway 4: Why Current Legal Guardrails are Failing

The Federal Rules of Evidence were designed for an era of physical forgeries and human-perceivable errors. Rule 901, governing authentication, currently sets a burden of proof – “plausibility” – that’s wholly inadequate for GenAI.

Judge Matthewman argues for a specific, burden-shifting solution to modernize this standard, including:

• Reasonable Suspicion: An opponent of the evidence must first meet a burden of “reasonable suspicion” or “probable cause” that the evidence is a deepfake.
• Preponderance of the Evidence: Once that threshold is met, the proponent of the AI evidence must then prove its authenticity by a “preponderance of the evidence.”

Judge Matthewman also views proposed Rule 707 as a misguided solution. By allowing machine-generated evidence without a human witness, it intentionally sidesteps cross-examination, which Judge Matthewman views as the “crucible of trial.”

Takeaway 5: Ten Core Components to Address the AI Onslaught

Perhaps the most notable takeaway is that Judge Matthewman proposed “ten core components for effectively dealing with the coming wave and onslaught of AI-generated evidence in our legal system 10 core components to restore the integrity of fact-finding”. Here’s a brief summary of those ten core components:

1. Mandatory Knowledge: Counsel must attend AI-specific training and CLEs.
2. Technical Proficiency: Lawyers must become adept at using illustrative aids while maintaining the ability to detect fakes.
3. Serious Sanctions: Severe sanctions for the reckless or intentional introduction of deepfakes.
4. Judicial Education: Comprehensive training for judges to understand the proper use, and the potential
5. misuse, of AI-generated evidence.
6. Amended Rules: Adoption of a “Brightline Rule” precluding AI evidence without a testifying witness (absent stipulation).
7. Pretrial Notice: Mandatory, early disclosure of intent to use AI-generated or enhanced evidence.
8. Robust Discovery: Mandatory access to AI provenance, including training data, metadata, hash values, and audit logs.
9. Pretrial Motion Practice: Mandatory “Liar’s Dividend” preclusion to prevent parties from making “baseless deepfake claims” without a good-faith basis.
10. Prompt Rulings: Judges must rule on AI challenges before trial to prevent “sandbagging.”
11. Updated Jury Instructions: Clear guidance on how to evaluate machine-generated reliability without assuming inherent objectivity.

Importantly, these core components require mandatory collaboration with ESI and technical experts. Technical proficiency is no longer a “niche” skill; it is an ethical imperative.

Judge Matthewman not only highlights key issues with the use and misuse of AI, but he also identifies meaningful steps toward a solution. It won’t be easy, which is perhaps why he concludes with the perspective of renowned Vietnamese Zen Buddhist monk, poet, and scholar, Thich Nhat Hanh: “Smile, breathe and go slowly.” Indeed.

So, what do you think? Are you prepared for the coming use and misuse of Artificial Intelligence in the courtroom? Please share any comments you might have or if you’d like to know more about a particular topic.

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.