A test comparing AI models found that Grok's model tends to select Republican positions and OpenAI's model tends to select Democratic positions when forced to choose between political stances.

"Test shows Grok's AI selects Republican positions, OpenAI's selects Democratic positions" — TechCrunch · AI & Technology

25 Oct 2026

View source

What it means

The TechCrunch article reports on a study by Convovo, a conversational AI research studio, that forced AI models to choose between political stances. The study found that , xAI's chatbot, recommended Republican candidates in roughly 92% of cases, while OpenAI's model recommended Democratic candidates in 97% of cases when forced to pick between candidates in hundreds of simulated races. This reveals clear partisan biases in the models' outputs.

Accuracy & sourcesGrounded

The claim comes from a study conducted by Convovo, a research studio led by a Cornell professor, and is reported by TechCrunch, a reputable tech news outlet. The specific numbers (92% for Grok, 97% for GPT) are cited, but the methodology and sample size are not detailed in the highlight; the reporting appears solid but is a secondary account of the primary study.

A test comparing AI models found that Grok's model tends to select Republican positions and OpenAI's model tends to select Democratic positions when forced to choose between political stances.

What it means

Accuracy & sourcesGrounded

How new

Go deeper