ChatGPT-using lawyer punished for citing fake testimony from made-up witnesses

The New Mexico Supreme Court held attorney Stephen Aarons in direct contempt after he filed a ChatGPT-generated appellate brief containing fabricated witness testimony, including fake police testimony, and other factual and legal errors. Aarons admitted he did not verify the AI-generated claims or authorities before filing and did not disclose the failures to his client; the court referred him to a disciplinary board. The case underscores legal, governance, and reputational risks from unverified generative-AI use in professional services.

Analysis

This is not a material earnings event for AI platforms, but it reinforces a commercially important distinction between consumer-grade model access and auditable professional workflow software. The beneficiaries over 6-18 months are likely legal-information incumbents RELX and TRI, whose proprietary-corpus, citation-validation, and enterprise indemnification capabilities can support premium pricing as firms formalize AI-use policies. Legal AI vendors that can document source provenance may gain share from ungoverned public-model usage, while raw-model providers face a modest reputational and procurement headwind rather than meaningful near-term financial liability.

The second-order effect is higher compliance spend at law firms, insurers, and corporate legal departments: mandatory human review, retrieval-augmented generation, and audit trails raise switching costs and favor integrated workflow vendors over standalone chat interfaces. Consensus may overstate the implication for broad AI adoption; professional users are unlikely to abandon AI, but will migrate toward constrained, source-linked deployments. The relevant 1-3 month catalyst is disclosure of tightened AI policies or product-demand commentary from RELX/TRI, while the thesis is falsified if legal customers continue choosing low-cost generic tools without paying for verification layers.

Key Decisions for Investors

• No directional trade in broad AI proxies from this event alone; the direct revenue impact is immaterial and headline-driven weakness in MSFT, GOOGL, or ORCL would be more likely a buying opportunity than a fundamental short signal.
• Add RELX and TRI to a 6-18 month watchlist for enterprise legal-AI monetization; initiate only if management quantifies legal workflow AI bookings, seat-price uplift, or retention improvement. A relative-long versus a basket of unprofitable legal-tech/private-market proxies is preferable to outright momentum exposure.
• Monitor upcoming RELX and TRI earnings for evidence that verification, citation, and workflow products are accelerating paid adoption. If AI-related investment rises without corresponding recurring-revenue growth or margin resilience, avoid the long thesis because implementation costs could absorb the pricing benefit.