It turns out AI is just like people. It has a hard time saying IDK.
I found this out from asking the cloud a question about “Salt,” a 2010 thriller starring Angelina Jolie and Liev Schreiber. Jolie plays a good agent, Schreiber a bad one who wants to kill nine million Arabs and Persians so the world will blame the U.S.
In the process, before Jolie saves the day, he knocks the president unconscious. While authorities suspect Jolie, it’s obvious that when the president awakens, he’ll finger Schreiber. But Jolie acts as though she’ll still be in trouble. At the end, she escapes and is shown stumbling through the forest in search of the sequel that that hasn’t yet occurred.
If you’re like me, you type these kinds of conundrums into your browser. What is the meaning of life? Do cats really dislike humans? I often end up at Quora, a Q&A website based in Silicon Valley.
When I browsed through the hole in “Salt”’s plot, I got an answer attributed to ChatGPT, Sam Altman’s chatbot — which obviously didn’t even know the plot of the movie but answered with supreme confidence nevertheless. Western civilization remains secure if AI has only evolved to the level of a precocious sophomore who’s called on in class when they haven’t done the reading.