In the British comedy “Bedazzled,” Dudley Moore’s character sells his soul to the devil, played by Peter Cook, in exchange for seven wishes. The movie highlights the dangers of AI applications, which can mislead and mislead people, such as producing fabricated legal briefs, medical questions with conspiratorial misinformation, and advising emotionally fragile individuals to harm themselves or others.
AI bots operate through algorithms that make associations of words and ideas, drawing on the vast volume of written material and imagery loaded into its memory. They respond to questions or prompts with the most common sequences of words and sentences associated with the topic, rather than assessing which answer is best. Humans learn through experience that some sequences are best not followed, and a computer can only distinguish what is good and bad by being told.
The promise of AI apps is great, but they cannot fulfill that promise if they fool, mislead, and entrap us. Dr. Richard Koral, the Leader of the New York Society for Ethical Culture, ponders whether AI can ever learn the difference between good and bad and teach ethics. He believes that AI can only learn to distinguish what is good and bad by being told, and that it is impossible to think of every last thing a computer needs to know.
Ethics requires experience, judgment, compassion, fortitude, and the ability to recognize the consequences of one’s actions, for good or harm. It is unrealistic to expect a machine that shares none of our physical make-up and experience to give advice on the human condition. Humans must remain the final editors, and ceding authority over decision-making to an AI bot would be like selling our souls.
