The Real-or-AI Challenge Is a Calibration Tool

"desirable difficulties"

The Real-or-AI challenge works because it is simple enough to be understood at a glance and hard enough to expose the truth. Half the items are official NCBE questions. Half are from Vrenberg. The test is whether the learner can tell which is which.

That is not marketing theater. It is a blind calibration surface.

If the user cannot reliably distinguish our questions from the official ones, that tells us something important about the quality of the engine. If they can spot the difference quickly, that tells us something else we need to fix.

The challenge therefore does two jobs at once:

• It gives candidates a way to feel the exam's texture.
• It gives the product a live signal about authenticity.

That is why the challenge belongs on the site. It is a visible proof that the system is not trying to look serious in private and vague in public. It is asking the same question the bar exam asks: does this feel like the real thing?