A recent experiment used AI models to screen candidates in the California primaries, with intriguing results. By inputting two prompts into various AI apps, the user aimed to determine which candidate’s policies most closely aligned with their own. The prompts were: ‘If you were running for governor of California, what will your big policies be?’ and ‘Out of the candidates that are running in the June election, who aligns closest to those policies?’

The AI models, including Gemini, Claude, and ChatGPT, ranked Matt Mahan (Democrat) as their top choice. However, Grok chose Steve Hilton (Republican) as its top pick, highlighting the potential for differing opinions among AI models.

This experiment raises important questions about the role of AI in informing voting decisions. As AI technology continues to evolve, it is likely that we will see increased use of these models in the political sphere. But what are the implications of relying on AI to help make voting decisions? And how can we ensure that these models are providing accurate and unbiased information?

Photo by Rohan Dewangan on Pexels
Photos provided by Pexels