Scope of this briefing: Note: requires attorney verification, given the developing and jurisdiction-specific nature of AI defamation liability.
On 24 July 2026, Delaware Superior Court Judge Meghan A. Adams denied Google's motion to dismiss defamation claims brought by Robby Starbuck over false statements generated by Google's Bard/Gemini AI systems (Starbuck v. Google LLC, No. N25C-10-211) Reason (Volokh Conspiracy), July 24, 2026.
What the Court Held
- Google argued the outputs were mere “hallucinations” lacking the characteristics of a published statement; the court declined to adopt that framing at the motion-to-dismiss stage.
- The court allowed Starbuck's claim that Google's AI exhibited a “deliberate, engineered” bias to proceed to discovery, again declining to resolve the factual dispute pre-discovery.
- The ruling does not decide the merits -- it only allows the case to proceed -- but it signals that “it's just an AI hallucination” is not, by itself, a viable basis to dismiss a defamation claim at the pleading stage.
Why It Matters for Clients
Courts are increasingly willing to treat AI-chatbot output as a “published statement” capable of supporting defamation liability -- an emerging legal theory with implications for any client operating or deploying consumer-facing generative AI tools, including customer-service bots and internal tools with external-facing outputs.
Practical Note
This is a single trial-court ruling on a motion to dismiss -- not binding precedent and not a final judgment. It is an early, non-final signal, not a settled rule.
CIDAH IN PRACTICE
CIDAH's own AI-disclosure footer, and its human-approval gate before any external send, exist precisely to prevent unreviewed AI output from reaching a client or the public unexamined -- this ruling is a live illustration of the exposure that gate is designed to avoid.
This document combines the firm's advanced AI system with experienced legal expertise, under the close supervision and approval of senior attorney.