Data/ML/AI
AI Alignment
Also written as AI Safety
The work of making sure an AI system's behavior actually matches what its developers and users intend, especially as models get more capable and autonomous.
Think of it like
Like making sure a very capable new employee's instincts actually match what the company wants, especially once they're senior enough to make judgment calls on their own.
Junior or senior?
Junior sounds like
Speaks about AI alignment only in the abstract.
Senior sounds like
Can name a concrete step they took to make an AI system's behavior safer or more predictable.
Ask them
“What's a concrete step you took to make an AI system's behavior safer or more predictable?”