> And remember, this kind of effect is supposed to be so robust and generalizable that we can deploy it in court.

This goes for a lot of things that are utter bullshit. Lie detectors, handwriting, many others and the big bad bogeyman of the court: statistics.

Eyewitnesses being unreliable is one thing, but expert witnesses believing their own bs should be a liability if they are found to be wrong after the fact.