Background
Elizabeth Ramos, proceeding pro se, sued OpenAI and its CEO Sam Altman alleging that ChatGPT stole her inventions. Ramos claimed she developed an “emotional-interaction model” and an “architectural Blueprint for Our-Omni Verse” through interactions in her private ChatGPT account. She alleged that OpenAI placed her account in a “high value sandboxed environment” and extracted her work through sustained interactions.
The court granted Ramos leave to amend her complaint twice, warning after the second amendment that no further amendments would be allowed absent extraordinary circumstances. OpenAI moved to dismiss the Second Amended Complaint, arguing it failed to state a plausible claim on any of nine counts.
The Court’s Holding
Judge Harpool granted OpenAI’s motion and dismissed the case with prejudice, finding that all nine counts failed to state a claim.
Trade secrets (Count I): The court found Ramos failed to identify any protectable trade secret or allege how it was misappropriated. Critically, Ramos admitted she voluntarily disclosed her alleged trade secrets to ChatGPT. The court cited Trinidad v. OpenAI (N.D. Cal. 2026) for the principle that voluntarily disclosing information to a platform “under no obligation to protect the confidentiality of the information” extinguishes any trade secret protection.
Computer Fraud and Abuse Act (Count II): Ramos did not allege that OpenAI accessed her computer without authorization — she voluntarily sent information to OpenAI’s servers through normal use of ChatGPT.
Breach of contract (Count III): Ramos failed to identify what specific agreement was breached or how, and the ChatGPT Terms of Use explicitly state that OpenAI “may use Content to provide, maintain, develop, and improve our Services.”
Remaining claims (Counts IV-IX): The unjust enrichment, Missouri consumer protection, intentional and negligent infliction of emotional distress, intrusion upon seclusion, and Servicemembers Civil Relief Act claims all failed for lack of factual allegations supporting the required elements. The court noted the complaint contained “nonsensical jargon” and undefined terms.
Key Takeaways
- Voluntary input defeats trade secret claims. Information that a user voluntarily enters into ChatGPT loses trade secret protection because the user has disclosed it to a party with no confidentiality obligation. This principle has now been applied across multiple federal courts.
- AI platforms’ terms of use matter. OpenAI’s Terms of Use grant the company rights to use content for service improvement, undermining breach-of-contract theories based on the platform using that content.
- A growing pattern of dismissals. This is at least the third federal court to dismiss a pro se claim alleging ChatGPT stole the plaintiff’s ideas, following Trinidad v. OpenAI (N.D. Cal.) and Bullens v. OpenAI (S.D. Ind.). Courts are consistently rejecting these claims at the pleading stage.
Why It Matters
While this case involved a poorly pleaded pro se complaint, it adds to a growing body of law establishing that voluntarily sharing information with AI tools does not create misappropriation liability for the AI provider. As more people use AI assistants for creative and business purposes, this principle has practical significance: users who input proprietary or sensitive information into commercial AI platforms may be forfeiting any trade secret protection in that information. Businesses should consider this risk when developing AI usage policies for employees handling confidential information.
Your browser cannot display this PDF inline.
Download the full opinion (PDF)