Study finds Grok and ChatGPT recommend opposing political candidates

· Tech · United States

A new study found that Grok, the AI model from Elon Musk's xAI, overwhelmingly recommends Republican candidates, while OpenAI's GPT favors Democrats, TechCrunch reported. The two models disagreed in more than 90% of nearly 500 election races examined. The findings suggest that model choice can meaningfully shape the political information users receive. The result adds to a growing body of research on political bias in large language models, an issue that has drawn scrutiny from regulators and lawmakers as AI assistants become a common source of information. For enterprises deploying AI, the study raises questions about how model outputs are evaluated and governed, particularly in politically sensitive contexts.

Why it matters：As AI assistants become a default source of information, evidence that model choice skews political recommendations will intensify pressure on developers and regulators to test and disclose bias.

Key points

  • Grok, built by Elon Musk's xAI, overwhelmingly favored Republican candidates.
  • OpenAI's GPT favored Democrats, per the study.
  • The models disagreed in more than 90% of nearly 500 races examined.

Sources (original reporting)

More stories from this day