Google is making major changes to how its Gemini AI handles mental health crises. The company announced new safety features that will connect users more directly to professional help when conversations suggest someone might be in danger.
The tech giant is also putting $30 million behind the effort over three years to help crisis hotlines around the world expand their services. Mental health affects over one billion people globally, and Google says it wants its AI tools to play a positive role rather than create new problems.
How does it work?
When Gemini spots a conversation that might indicate suicide risk or self-harm, it will now show a simplified ‘one-touch’ interface. This connects users directly to crisis hotline resources where they can:
- Start a chat session
- Make a phone call
- Send a text message
- Visit the crisis hotline website
The system keeps these help options visible for the rest of the conversation once activated. Google developed this with clinical experts and says it’s designed to encourage people to actually seek help.
For less urgent mental health discussions, Gemini will show an updated ‘Help is available’ module that points users toward appropriate resources and information.
Why does it matter?
More people are having complex, personal conversations with AI chatbots like Gemini. This includes discussions about mental health crises, which puts pressure on companies to handle these situations responsibly.
Google admits that AI tools can create new challenges for mental health support. But as these tools become part of daily life for millions of users, the company believes they need to connect people to real human help when it matters most.
The $30 million in funding will help crisis hotlines scale up their capacity to handle more people reaching out for immediate support. Google is also expanding its partnership with ReflexAI, including $4 million in direct funding to help organizations train their mental health support staff using AI-powered simulations.
The context
Google has specific protections for younger users of Gemini, including responses designed to avoid harmful topics. The company’s clinical, engineering, and safety teams have been training the AI model to recognize when someone might be in an acute mental health situation.
The company is clear that Gemini isn’t a replacement for professional therapy or crisis support. Instead, it’s trying to make the AI better at recognizing when users need real help and connecting them to it quickly.
These changes are part of Google’s longer-term effort to use its technology responsibly in mental health situations. The company says it wants to make support more accessible and effective, while creating a safer digital environment for people to explore and learn.