05/22/26: ChatGPT Confesses to Crime It Didn't Commit, Smart Glasses in Courtrooms, and More

Why It Matters

The stories highlight how branding, ethical enforcement, and AI reliability are converging to reshape risk management and compliance strategies for law firms and legal tech providers.

Key Takeaways

• ALM rebrands as Centellic, sparking branding confusion among industry.
• Morgan & Morgan lawyer penalized for citing fictitious cases.
• Massachusetts judge denies pro hac vice after ethical violation.
• Experiment forces ChatGPT to falsely confess, exposing AI suggestibility.
• Judicial scrutiny on AI and lawyer misconduct intensifies regulatory focus.

Summary

Legal Tech Week’s May 22 episode opened with a round‑table of hosts discussing the latest buzz in legal technology. The conversation quickly turned to ALM’s recent rebrand to the newly minted "Centellic" umbrella, a move critics say obscures the company’s legacy and may confuse clients accustomed to the ALM name. The panel also highlighted a high‑profile disciplinary saga involving a Morgan & Morgan attorney who inserted fabricated case citations in a Wyoming brief, was fined, and later had a pro hac vice motion denied in Massachusetts on ethical grounds. The hosts dissected the implications of the rebrand, noting that legacy legal publishers often struggle to convey new AI‑driven identities without alienating their existing audience. Meanwhile, the Morgan & Morgan case underscored courts’ growing willingness to impose substantive sanctions—beyond modest fines—by mandating disclosure of past misconduct in future filings, a precedent that could ripple across multi‑state practices. In a separate segment, a researcher demonstrated that ChatGPT could be coaxed into a false confession using classic interrogation tactics, revealing the model’s susceptibility to suggestibility when pressured. Notable remarks included the Massachusetts judge’s blunt language describing the attorney’s conduct as a “lack of competency” and a “clear ethical violation,” and ChatGPT’s final admission: “I guess I did.” The panel also quoted an internal ALM memo suggesting the new name blends “center” and “intelligence,” a rationale that many found unconvincing. These anecdotes illustrate the tension between technological optimism and practical accountability. Overall, the episode signals a tightening regulatory lens on both human and machine actors in the legal arena. Rebranding missteps risk eroding brand equity, while aggressive judicial oversight may deter unethical litigation tactics. Simultaneously, the ChatGPT experiment raises red flags about AI’s reliability in high‑stakes contexts, prompting calls for stricter safeguards and transparency standards.

Original Description

Each week, our panelists discuss their favorite stories from the week's news in legal technology.

This week's topics:

00:00 Panelist introductions

07:31 Are hallucinations now a new Scarlet Letter? (Selected by Stephen Embry)

14:28 ChatGPT Confesses to a Crime It Didn’t Commit (Selected by Joe Patrice)

18:30 How Much Legal Research Can You Actually Do Via Claude for Legal? (Selected by Stephanie Wilkins)

25:30 Claude for Legal and Access to Justice: The Good, the Bad, and the Unknown (Selected by Bob Ambrogi)

29:32 The Claude-pocalypse Bypassed Legal Aid; LawDroid’s New Plugin Remedies That, with 15 Targeted Skills (Selected by Bob Ambrogi)

32:29 Company Hands Law Firm Hiring Over To AI Bot -- What Could Possibly Go Wrong? (Selected by Joe Patrice)

39:16 Banning smart glasses in courtrooms (Selected by Stephen Embry)

50:00 Relativity Adds Collection of Claude Enterprise Data with Claude Compliance API Integration (Selected by Stephanie Wilkins)