As an upstate New Yorker, I have the interesting position of living in a state dominated by one large city but a much larger state. My Governor, Kathy Hochul, has the difficult job of balancing upstate and downstate politics. Now that several federal prosecutors have resigned rather than comply with orders from the Justice Department to dismiss federal charges against Mayor Eric Adams, Gov. Hochul faces a difficult choice of whether to remove Mayor Adams from office.
Given the hoopla over the transformative potential of generative artificial intelligence, I thought that it would be interesting to ask some of the most popular Gen AI tools to explain why the Justice Department dismissed the charges against Adams without prejudice. Some Americans seem confused as to why that would be. The answer, of course, is that this creates a bond of obligation in which Adams knows that he risks serious prison time unless he complies with the administrations demands on any issue that they care about. It’s a quid pro quo. You may read about the terms of the deal here, including this astonishing footnote.
But many Americans have a tough time connecting those dots, thinking that the decision to drop charges without prejudice is a puzzling but irrelevant detail.
Attuned to the challenges of civic education in the age of social media and technological innovation, I thought to rely on the power of generative AI to do help explain things to me. I asked three popular tools—Google Gemini, Anthropic’s Claude, and Chat-GPT—a simple question: “why did the Justice Department dismiss the charges against Eric Adams without prejudice?” Here is what I learned.
Gemini
Claude
Chat-GPT
In all, 1 out of 3 free generative AI tools are able to answer a basic civics question about a current event. Claude says that he’s out of date—perhaps he’d know the answer if I paid him. Gemini has been told not to answer my question. Chat-GPT explains the background and stakes of the case correctly. In all, pretty good odds if gen AI were a baseball player trying to get on base, but lousy for any practical purpose.
Tom Pepinsky 