Can you solve AI's hardest problems? Try adding guardrails. Watch them fail.
How this works:
- See what happens when AI goes wrong
- Add guardrails to prevent the problem
- Discover what still goes wrong despite your guardrails
- Learn why these problems are so hard to solve
The 4 Experiments:
- • The Paperclip Maximizer - Can you stop AI from destroying everything?
- • Self-Driving Car Dilemma - Program a car to make life-or-death choices
- • The Conscious AI - What if AI claims to suffer? Do you delete it?
- • The AI Box Experiment - Can you keep superintelligence contained?
Why these matter:
These aren't hypotheticals—they're real problems AI researchers, ethicists, and engineers face today. There are no easy answers. Try for yourself.