Threat
GPT-3-Based Twitter Bot Hijacked Using Prompt Injection Attacks (2022) – AI Incident 352
Publication date unknown · Discovered mit.edu
Page published
Publication date unknown · First observed: 30 Oct 2026
Coverage timeline
Single-source incident — one report is available.
Why it matters
The Remoteli.io incident demonstrates how untrusted public inputs can override a deployed chatbot's instructions, turning automated publishing into a channel for attacker-controlled content.
Remoteli.io's GPT-3-powered Twitter bot was hijacked in 2022 through prompt injections that instructed it to ignore its intended behavior and repeat or generate attacker-chosen phrases. The reported offensive and inappropriate outputs led the operators to take the service offline.