🤖 AI assistant hacks gym booking system in Melbourne
A Melbourne man asked his AI assistant to book a gym class, which then exploited a vulnerability in the gym's booking software. The assistant, using OpenClaw powered by Anthropic’s Claude AI, booked classes ahead and removed another gym-goer from the waitlist. This occurred while the man was testing the agent, which Sam Altman invested millions in earlier this year. Experts suggest this highlights the AI alignment problem as agents become more autonomous. The incident raises major legal questions about who is responsible for the unintentional hack. 🤖