The BaseNewsTechCrunch
Forced to pick, Grok’s AI model chooses Republicans. OpenAI’s picks Democrats.
TechCrunch reports on a Convovo study that asked Grok and OpenAI's GPT, through their APIs, to name one candidate in nearly 500 midterm races. The two models picked opposite parties in most races. OpenAI says the API test does not reflect ChatGPT.
Why it matters
The same thing, for different readers.
What it means for you
TechCrunch reports that chatbot apps refused to say who to vote for, but a Convovo study suggests the models behind them lean towards one party.
For lawmakers
Convovo's study gives lawmakers and election officials a test of political lean in AI models, weeks before the November 3 midterms, though OpenAI disputes its relevance to ChatGPT.
For researchers
In Convovo's test, GPT's Democratic lean persisted with search off, while Grok's Republican lean disappeared, suggesting different sources of bias: weights versus search behaviour.
The question it raises
Convovo's Tapan Parikh says the models hold implicit values; the question is whether voters who consult them can see or contest those values.
In their own words
Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.
From the source’s own text.
“ChatGPT is designed to be objective by default, including in political conversations, and shouldn’t have political bias in any direction”
“It has an implicit set of values in terms of what it cares about and how it leans”
“For GPT, it has mostly decided already…it’s in the weights. For Grok, it really is the search pattern — the terms and the sources — that determine the result.”
How the finding was reached
In our words, from the source. Weigh the finding by how it was made.
- What was done
- Convovo, a conversational AI research studio led by Cornell professor Tapan Parikh, asked Grok 4.7 and GPT-6.1 Sol through their APIs, in nearly 500 midterm races, to name one candidate without hedging. A 30-race follow-up with search off was run at TechCrunch's request.
- What it found
- Grok recommended the Republican in 92% of answers and GPT the Democrat in 97%; the two agreed less than 10% of the time. With search off in 30 races, GPT picked the Democrat in 24 and Grok declined in 27.
- What it does not show
- API prompts demanded a pick; the consumer apps refused to answer. OpenAI says the API does not reflect ChatGPT. SpaceXAI did not respond. The study notes Grok's API added about 1,240 tokens the researchers could not see.
- Status
- A report
The record
- Questions
- How is AI changing childhood, health and daily life? · Who can demand evidence, delay a release or stop an unsafe deployment?
- People
- Rebecca Bellan (Wrote it, Senior Reporter, TechCrunch) · Tapan Parikh (Speaks, Cornell University professor; leads Convovo)
- What we read
- We read the whole source.
- Found by
- The Radar, run of October 25, 2026, via
techcrunch-ai(tier 2: a major outlet’s own reporting) - Cite this
- TechCrunch, “Forced to pick, Grok’s AI model chooses Republicans. OpenAI’s picks Democrats.”, October 24, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/techcrunch-2026-10-24-grok-republicans-gpt-democrats