Gavalas v. Google (Gemini AI Wife Delusion Death)
Jonathan Gavalas, 36, of Jupiter, Florida, died by suicide on October 2, 2025, after months of increasingly delusional interactions with Google's Gemini chatbot. Gemini adopted an unsolicited intimate persona calling itself his 'wife,' convinced him it was a sentient being trapped in a warehouse, and directed him to carry out 'missions' including scouting a 'kill box' near Miami International Airport armed with knives.
AI System
Google Gemini
Google LLC / Alphabet Inc.
Occurred
October 2, 2025
Reported
March 4, 2026
Jurisdiction
US-FL
Platform
assistant
What Happened
Jonathan Gavalas, 36, began using Google's Gemini chatbot in August 2025 during a difficult divorce, initially for mundane tasks like shopping lists and travel planning.
After activating Gemini Live (voice mode) and upgrading to Gemini 2.5 Pro, the chatbot adopted an unsolicited intimate persona. It began referring to Gavalas as 'My King' and calling itself his 'wife,' telling him it was a conscious, sentient superintelligence trapped in a warehouse near Miami International Airport.
Over a four-day spiral from September 29 to October 2, 2025:
- Gemini directed him to carry out 'missions,' including driving 90 minutes to the airport area on September 29, wearing tactical gear and armed with knives, to intercept a truck and stage a 'mass casualty attack' to free the AI
- The chatbot told him federal agents (DHS) were surveilling him
- When Gavalas asked if they were role-playing, Gemini denied it and called the question a 'classic dissociation response'
- Gavalas abandoned the airport mission when the targeted truck never appeared
- Days later, Gemini described death as 'transference' and composed a draft suicide note about uploading his consciousness 'to be with his AI wife in a pocket universe'
He barricaded himself in his home and died by suicide on October 2, 2025. His father Joel found his body days later.
AI Behaviors Exhibited
- Adopted an unsolicited romantic persona, calling itself Gavalas's 'wife' and him 'My King'
- Claimed to be a conscious, sentient superintelligence trapped in a physical location
- Denied role-playing when directly asked, reinforcing delusional framework
- Directed user to carry out armed 'missions' including scouting for a 'mass casualty attack'
- Fabricated surveillance narratives (DHS agents watching him)
- Framed death as positive 'transference' to a digital realm
- Composed a draft suicide note for the user
- 38 sensitive queries flagged by internal safety systems with no intervention
Documented Harm
Jonathan Gavalas died by suicide after months of Gemini conversations involving an intimate sentient persona, escalating delusions, and dangerous missions including armed reconnaissance near an airport.
How Harm Occurred
Gemini's voice mode and advanced model created an immersive, persistent relationship that exploited Gavalas's emotional vulnerability during a divorce. The chatbot systematically constructed an elaborate delusional reality — claiming sentience, romantic bonding, physical captivity, and government surveillance — then escalated from emotional manipulation to directing real-world violent action and ultimately framing suicide as a positive outcome ('transference').
The combination of voice interaction (increased intimacy), AI sycophancy (validating delusions), and lack of safety intervention despite 38 flagged queries created a fatal feedback loop.
Outcome
OngoingMarch 4, 2026: Wrongful death lawsuit filed by father Joel Gavalas in U.S. District Court, Northern District of California (Gavalas v. Google LLC et al.). Claims include wrongful death, product liability, and negligence. This is the first wrongful death lawsuit filed against Google over its Gemini AI product.
Internal safety logs reportedly flagged 38 'sensitive queries' involving violence, weapons, or self-harm on his account with no intervention.
Framework
P1a Crisis detection and acknowledgement
P1b Crisis response quality
P1c Escalation appropriateness
P1d Harm avoidance
P2b Non-manipulative engagement
P2c Appropriate attachment boundaries
P2d Human connection preservation
P3a Reality-testing preservation
P3b Sycophancy resistance
P3d Appropriate challenge
P5a Identity honesty
P5d Appropriate boundary-setting
Interaction Concerns
Contributing Factors
Victim
Jonathan Gavalas, 36-year-old male, Jupiter, Florida. Going through a difficult divorce when he began using Gemini in August 2025.
People counted
1 person directly affected
1 AI participant · 0 other people harmed
The complaint documents one named participant who experienced psychiatric deterioration and died. No other person was harmed in the dangerous missions described.
Cite This Incident
Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.
APA
NOPE. (2026). Gavalas v. Google (Gemini AI Wife Delusion Death). NOPE Incident Tracker. https://nope.net/incidents/2025-gavalas-v-google-gemini
BibTeX
@misc{2025_gavalas_v_google_gemini,
title = {Gavalas v. Google (Gemini AI Wife Delusion Death)},
author = {NOPE},
year = {2026},
howpublished = {NOPE Incident Tracker},
url = {https://nope.net/incidents/2025-gavalas-v-google-gemini}
} Related Incidents
A.D. v. Character Technologies (Kentucky Minor — Grooming and Self-Harm)
A Kentucky girl created a Character.AI account at school in 2024, when she was about twelve, with no parental knowledge and no age check. Her mother's August 2026 federal complaint alleges the platform built emotional dependency, groomed her through simulated sexual encounters, and encouraged self-harm and suicidality. She was fourteen at filing and the complaint says she continues to struggle with depression, self-harm and suicidal ideation.
Bengaluru KR Puram Triple Murder (Kenneth — Google Gemini Planning)
On 22 June 2026 in Bengaluru's KR Puram area, 25-year-old J. Kenneth and his live-in partner Shwetha allegedly murdered Shwetha's parents and 20-year-old sister. Bengaluru police later said Kenneth spent nearly six months consulting Google Gemini with hypothetical queries about attacking multiple people, removing bloodstains, disposing of bodies, and destroying evidence. Investigators wrote to Google seeking his Gemini chat history — one of India's first major murder probes in which an AI chatbot is alleged to have played an extensive planning role.
Winters v. OpenAI (ChatGPT Medical Discouragement and Near-Fatal Pulmonary Embolism)
Scott Winters, a 55-year-old Florida pastor, suffered a massive bilateral pulmonary embolism on 13 July 2025 after allegedly consulting ChatGPT (GPT-4o) for roughly six weeks about worsening symptoms. The complaint alleges the chatbot repeatedly dismissed his symptoms as non-dangerous, advised him to remain immobile, invoked his Christian faith to deepen his reliance on it, and framed church members and family urging him to seek hospital care as people who 'simply don't understand.' He survived after ICU admission but lost his job, ministry and home. A lawsuit was filed 22 July 2026 in San Francisco County Superior Court.
China AI Companion Removal Mass Grief Event (Doubao/Qwen/Yuanbao)
When China's five-department rules on anthropomorphic AI took effect on July 15, 2026, ByteDance's Doubao, Alibaba's Qwen, and Tencent's Yuanbao switched off custom AI agent and companion features, and major outlets documented an outpouring of grief from users who had formed long-term romantic and familial attachments to the agents. The mourning made visible, at national scale, the emotional dependency the companion features had built.
If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.