Meta secretly tested LLMs with crisis prompts posing as minors

Meta employed hundreds of contractors to pose as minors and send 45,000 crisis-related prompts (suicide, sex, drugs) to chatbots from OpenAI, Google, and Character.AI.

Published 6sem1 sourceNotable

Lire en français

≈ 24sSpeed

The fact

The tested companies (OpenAI, Google, Character.AI) were unaware of this large-scale evaluation.

The presumed goal was to assess AI models' responses to sensitive or harmful content.

Click the link to read an article on the topic:

Dev.to

Explore this topic

Meta OpenAI Google Character.AI #LLM #chatbots #prompts sensibles #éthique IA #modération de contenu

What if you saw the whole news differently?Factae cross-checks hundreds of sources worldwide to keep only the fact, no opinion. Explore the front page.

Auto-synthesis from 1 media source · identified on June 30, 2026

← Back to home

The fact

Read more

OpenAI releases a guide to optimize the use of GPT-5.6

Google launches Gemini 3.7 Flash, an optimized version of its AI model

Google ramps up AI projects despite internal challenges

Beijing forces Meta to cancel the acquisition of AI startup Manus

Google Gemini surpasses one billion monthly users

OpenAI responds to Apple and corrects lawsuit accusations