Dan Hendrycks, director of the Center for AI Safety, also pointed out that while current AI apps operate using enormous stores of data, the tech currently lacks the trial-and-error correct-and-improve ...