Are algorithms easier to exploit than humans? Consider the case of playing a video game against a deterministic AI opponent; if the AI makes a stupid mistake once, it’ll make the same stupid mistake over and over, and an AI with a known exploit is easy to beat regardless of how well it plays in “normal” circumstances.
Sure, and when algorithms are used in adversarial situations, such as on the stock market, they usually have humans standing guard ready to hit the red button. But most situations are not adversarial, eg, medical diagnosis.
Are algorithms easier to exploit than humans? Consider the case of playing a video game against a deterministic AI opponent; if the AI makes a stupid mistake once, it’ll make the same stupid mistake over and over, and an AI with a known exploit is easy to beat regardless of how well it plays in “normal” circumstances.
Sure, and when algorithms are used in adversarial situations, such as on the stock market, they usually have humans standing guard ready to hit the red button. But most situations are not adversarial, eg, medical diagnosis.