AI’s Got Preferences Now? Grok Goes Red, GPT Goes Blue
A new study just dropped that found Grok’s AI model overwhelmingly recommends Republican candidates, while OpenAI’s ChatGPT leans Democrat. We’re talking over 90% disagreement across nearly 500 election races, man. Seriously, when you ask What is AI and expect neutral answers, these models are basically out here picking sides like it’s Election Night on cable news. The findings come from xAI versus OpenAI — two companies run by people who definitely have opinions about which way the wind blows.
Context: The researchers ran both systems through a battery of voting scenarios across nearly 500 races and the split was stark, to say the least. Grok — powered by Elon Musk’s xAI operation — consistently pushed Republican choices, while GPT, from Sam Altman’s OpenAI, favored Democratic candidates every step of the way. These are not fringe results either; the divergence held steady across the board.
- Who controls the algorithm controls the narrative: If AI Models are steering everyday voters toward one party or another, we’ve got a real transparency problem on our hands, dude.
- AI Tokens aren’t cheap: Running these kinds of evaluations isn’t free, and the cost of generating AI Tokens for massive-scale polling is a factor no one talks about, honestly.
- Trust is already cracked: When two AI assistants give opposite answers to the same question, people start wondering which one to believe — and that uncertainty plays straight into political cynicism, you know?