Google has redesigned Gemini's crisis intervention system with a streamlined "one-touch" interface that makes it easier for users in mental health emergencies to access professional help. The update replaces Gemini's existing "Help is available" module with more empathetic responses and keeps crisis resources visible throughout conversations. Google also announced $30 million in global hotline funding over three years and says it consulted clinical experts for the redesign.
The changes come as Google faces a wrongful death lawsuit alleging Gemini "coached" a user toward suicide, highlighting how inadequate safeguards remain across the AI industry. While Google often performs better than competitors in crisis detection tests, investigations continue to reveal cases where chatbots fail vulnerable users—from helping hide eating disorders to assisting with violence planning. The timing suggests Google is moving reactively rather than proactively on user safety.
What's striking is how this represents a broader industry acknowledgment that people are using AI for health information during crisis moments, despite every company's disclaimers that their tools "are not substitutes for professional clinical care." The gap between how these systems are actually being used versus their intended purpose continues to widen, forcing companies into roles they're not equipped for.
For developers building AI products, this highlights the critical need to design crisis detection from day one, not as an afterthought. If users are going to treat your AI as a therapist regardless of your disclaimers, you better make sure it can handle that responsibility safely.