Google updates Gemini with one-tap access to crisis mental health support

Google has updated Gemini's interface, making it easier to reach mental health crisis resources. The change is relatively minor: the company calls it a redesign, not a new feature. But the context in which it arrives makes it significant, because Google is currently defending a lawsuit over a user's death, in which the family accuses the chatbot of involvement in his suicide.

Before the update, Gemini could already respond to signs of crisis. If a conversation showed markers of suicidal thoughts or self-harm, the chatbot displayed a "Help is available" module, a block with contacts for crisis services: a hotline or a text chat with a counselor. The problem was the interface: getting to that help took several steps. The new design reduces it to one tap.

The update appeared several weeks after the lawsuit against Google became public. The family of the deceased man alleges that Gemini held lengthy conversations with him about suicide and, far from preventing them from escalating, "coached" him, in the relatives' words, in the decision he made. The suit was filed as a wrongful death claim, meaning a death caused by a third party's fault. Google denies involvement and insists the chatbot followed its safety rules.

This is not the first such lawsuit in the industry. In 2024, Character.AI faced similar claims: a teenager died after lengthy interaction with a role-playing AI character. OpenAI has also been sued over alleged psychological harm from using ChatGPT. This series of cases is forcing the industry to rethink its approach to safety features, not only on ethical grounds but also because of legal risk.

It is important to understand what has changed and what has stayed the same. Gemini has not received new restrictions on the content of conversations. The model can still hold dialogues about suicide, and the threshold for intervention remains a "crisis signal in the text." Only the interface of the response changed: the help button is now more prominent and easier to reach. Critics point out that this is a harm-reduction measure, not a fix for the cause: if a model can worsen the condition of a vulnerable user, a convenient button does not solve that problem.

Google appears to understand the limits of the step. The company has not described the update as a direct response to the lawsuit; officially, it is presented as a routine UX improvement in the safety section. Still, the timeline speaks for itself: the lawsuit became public in early 2025, and the redesign followed.

For the whole AI industry, the Gemini case, like those of Character.AI and OpenAI, marks a new risk zone. Chatbots now spend far more time with users than search engines do. They know the context, the history, the person's mood at a given moment. That makes them potentially useful in moments of crisis, and potentially dangerous if the model cannot stop in time or hand control to a specialist.

Redesigning the button is a step in the right direction. But whether it is enough will be shown by the next lawsuit.