4 ms·

Is there a theorem/law of AI for this: Every AI system is at least vulnerable to data is has never seen before.