AI Safety Hackathon Budapest: Prototyping Alignment & Governance Tools (sponsored by Lovable)

​Build an AI safety tool in one afternoon.

​Most AI safety ideas stall in the gap between concept and working prototype. This hackathon closes it: in a single afternoon, you’ll build and ship a real tool tackling an alignment, monitoring, or governance problem — using Lovable’s vibe-coding platform to go from prompt to functioning app in minutes, no setup or boilerplate.

​Challenge Tracks

​🔍 Track A: Interpretability & Visualization
Attention explorers, feature inspectors

​🛡️ Track B: Constitutional AI & Monitoring
Real-time guardrails, red-teaming dashboards

​⚖️ Track C: Governance, Policy & Risk
EU AI Act compliance checkers, lab commitment trackers

​Agenda: Full schedule here

​What to bring: Laptop, charger, and an active Lovable account

​Perks: 100 Lovable credits per participant - plus, members of the top 3 voted teams each win a book!

​🗓️ Registration deadline: November 6th, 18:00 CET — secure your spot soon!