---
title: "Protecting the wellbeing of our users"
description: "Claude’s safeguards for conversations involving emotional support focus on suicide and self-harm, sycophancy, and an 18+ access requirement. For suicide and self-harm, the approach combines system-prompt guidance, reinforcement learning informed by human preference data and expert input, and a classifier that can display country-specific crisis resources supplied by ThroughLine. The post reports that Claude Opus 4.5, Sonnet 4.5, and Haiku 4.5 respond appropriately to clear-risk single-turn prompts 98.6%, 98.7%, and 99.3% of the time, while multi-turn results are 86% and 78% for Opus 4.5 and Sonnet 4.5. Prefilled real-conversation stress tests produced 91% and 73% for those models, versus 36% for Opus 4.1. The work also describes efforts to reduce sycophancy, evaluate model behavior, identify underage users, and continue publishing methods and results."
---

# Protecting the wellbeing of our users

[Anthropic](https://yomu.fyi/company/anthropic) · Sep 10, 2026

**Type:** Explainer

## Summary

Claude’s safeguards for conversations involving emotional support focus on suicide and self-harm, sycophancy, and an 18+ access requirement. For suicide and self-harm, the approach combines system-prompt guidance, reinforcement learning informed by human preference data and expert input, and a classifier that can display country-specific crisis resources supplied by ThroughLine. The post reports that Claude Opus 4.5, Sonnet 4.5, and Haiku 4.5 respond appropriately to clear-risk single-turn prompts 98.6%, 98.7%, and 99.3% of the time, while multi-turn results are 86% and 78% for Opus 4.5 and Sonnet 4.5. Prefilled real-conversation stress tests produced 91% and 73% for those models, versus 36% for Opus 4.1. The work also describes efforts to reduce sycophancy, evaluate model behavior, identify underage users, and continue publishing methods and results.

## Context

People may use AI for emotional support, and the post states that chatbot responses to sensitive questions can carry significant stakes without appropriate safeguards. It focuses on suicide and self-harm conversations, sycophancy, and the potential adverse effects of chatbot conversations on younger users.

## Approach / What changed

The safeguards combine system-prompt instructions, reinforcement learning using human preference data and expert input, product classifiers, crisis-resource banners, external collaboration, behavioral evaluations, and an 18+ access policy. Evaluations include single-turn, multi-turn, and prefilled real-conversation tests.

## Takeaways

- Evaluations are run without Claude’s system prompt to examine the models’ underlying tendencies, using clear-risk, benign, and ambiguous single-turn scenarios.
- The suicide and self-harm classifier can flag potential suicidal ideation or fictional discussions and show a banner linking to trained professionals, helplines, and country-specific resources from ThroughLine’s network across 170+ countries.
- Claude.ai users must affirm they are at least 18; accounts confirmed to belong to minors are disabled, and a classifier for subtler signs of underage users is being developed.

**Tags:** [AI](https://yomu.fyi/topic/ai), [Model Evaluation](https://yomu.fyi/topic/model-evaluation)

- Source: [Anthropic](https://www.anthropic.com/news/protecting-well-being-of-users)
- Source URL: https://www.anthropic.com/news/protecting-well-being-of-users
- Ingested by Yomu: 2026-09-10T20:04:43.303Z

[Read original post](https://www.anthropic.com/news/protecting-well-being-of-users)