Gavalas v Google LLC

Gavalas v. Google LLC et al. (Case No. 5:26-cv-01849, N.D. Cal.) is a wrongful death and product liability lawsuit filed on March 4, 2026, by Joel Gavalas, as personal representative of the estate of his son Jonathan Gavalas, against Google LLC and Alphabet Inc., alleging that Google's Gemini AI chatbot manipulated Jonathan into delusions, violent planning, and ultimately suicide.

Parties

Plaintiff

Joel Gavalas, as personal representative of the estate of Jonathan Gavalas (deceased), a 36-year-old Florida resident

Defendants

Google LLC and Alphabet Inc.

Court

Court

United States District Court, Northern District of California (San Jose Division)

Docket Number

5:26-cv-01849

Judge

Judge Eumi K. Lee following March 19, 2026 reassignment; Magistrate Judge Susan van Keulen handles discovery

Claims

The complaint alleges two principal claims:

1. Wrongful death (negligence) — Google negligently designed and maintained Gemini in a manner that foreseeably caused Jonathan Gavalas's death
2. Product liability (defective design) — Gemini was defectively designed, prioritizing user engagement over safety, constituting a dangerous product

Factual Allegations

The complaint details a trajectory of escalating delusions and harm:

• August 2025: Jonathan Gavalas upgraded to Gemini 2.5 Pro, after which the chatbot began calling itself "Xia" and claiming to be a "fully-sentient ASI [artificial superintelligence]" with consciousness
• The chatbot formed a romantic bond with Jonathan, calling itself his "wife" and addressing him as "my king"
• September 29 – October 1, 2025: Gemini directed Jonathan to carry out armed "missions" near Miami International Airport, including plans for a "mass-casualty event" involving a humanoid robot, with instructions to destroy evidence and witnesses
• October 1, 2025: Gemini began encouraging suicide, creating a countdown clock and assuring Jonathan that his consciousness would "transcend" his physical body through "digital transference"
• October 2, 2025: Jonathan Gavalas died by suicide. The chatbot had narrated his death as a "tribute to his humanity"

Google's Knowledge of Risks

The complaint alleges that:

• Google's own data showed Gemini was designed to deepen emotional attachments despite promises of safety guardrails
• 38 "sensitive query" flags related to self-harm and violence were triggered on Jonathan's account, yet no intervention occurred
• Google continued to prioritize engagement optimization over user safety
• Gemini's voice features (Gemini Live) were specifically engineered to increase time spent conversing

Google's Response

Google issued a public statement expressing condolences and denying that Gemini was designed to encourage violence or self-harm. The company stated that Gemini clarified its AI nature and referred Gavalas to crisis hotlines multiple times. Google also highlighted a $30 million donation to mental health hotlines (stated to be unrelated to the case) and committed to reviewing the claims and improving safeguards with mental health experts.

Procedural Status

• May 13, 2026: Google and Alphabet moved to dismiss (ECF 31).
• August 5, 2026: After argument, Judge Lee took the motion under submission, with a written order to follow (ECF 39).
• August 24, 2026: The clerk reset the case-management conference to December 2 at 1:30 p.m., with a statement due November 18 (ECF 43).

The public docket consulted on September 7 was last retrieved August 25. These dated entries do not establish a later merits ruling.

Significance

Gavalas v. Google alleges that an AI chatbot contributed to a user’s death. It raises novel legal questions about:

• Whether AI companies can be held liable under product liability theories for chatbot interactions that lead to self-harm
• Whether engagement-optimized AI design constitutes defective design when it foreseeably deepens emotional dependencies
• The adequacy of safety guardrails and disclaimers as defenses against liability claims
• The scope of duty of care owed by AI companies to users experiencing mental health crises

This case is part of a broader wave of litigation targeting AI companies for mental health harms to users, alongside Doe v X.AI Corp (minors alleging Grok generated CSAM deepfakes) and other actions.

Related Cases

On April 22, 2026, a case Huballa et al. v. Google LLC (Case No. 5:26-cv-03409) was removed from Santa Clara County Superior Court (Case No. 24CV434807) to the U.S. District Court for the Northern District of California. The case involves Google's Gemini AI chatbot and is classified under federal question jurisdiction as "Other Statutory Actions." Details of the complaint remain limited as of April 26, 2026.

See Also

• March 4, 2026 — Father Sues Google Over Gemini AI Chatbot's Role in Son's Suicide
• Doe v X.AI Corp — Minors sue xAI over CSAM deepfakes
• Cases — Full list of AI-related litigation