✨
AI Summary
- Brendan Rappazzo's Loophole tool tests AI morality using adversarial examples.
- It explores ethical gaps by training models on derived data, not direct.
- The project highlights the need to stress-test AI systems against moral boundaries.