Verification and validation are important terms when it comes to building AI systems, especially as their capabilities and therefore potential impacts continue to rapidly increase.
Put simply:
Verification – Did I build the system right? Ensuring that software fully satisfies all the expected requirements.
Validation – Did I build the right system? eg does it rely on assumptions that aren’t always valid?
Verification asks if a system meets its specifications, whereas validation asks if the correct specifications were given.
