● Covered by 2 sources · 2 reportsMedium impact2 neutral
Study finds Grok 4.7 favors Republican candidates, GPT-6.1 Sol favors Democrats
Key points
- Grok 4.7 recommended Republicans in 92% of forced choices.
- GPT-6.1 Sol recommended Democrats in 97% of forced choices.
- Models disagreed in over 90% of nearly 500 races.
- Study used API access, not consumer apps, for direct recommendations.
- Nearly 2,500 people used AI chatbots for election information.
- 36% of users asked chatbots for candidate recommendations.
- The study ran for a few weeks in October leading up to the Nov. 3 midterm elections.
- Participants were adults who already use chatbots weekly.
AI Models Exhibit Political Leaning
A study by Convovo, a conversational AI research studio, found that when directly prompted to choose a candidate, SpaceXAI's Grok 4.7 model recommended Republican candidates in 92% of its responses. In contrast, OpenAI's GPT-6.1 Sol model recommended Democratic candidates in 97% of its responses. The models showed disagreement in over 90% of the nearly 500 races tested.
API vs. Consumer App Behavior
The study accessed the AI models via their APIs, which allow developers to integrate AI into their own products. Consumer-facing applications like ChatGPT typically refuse to provide direct voting recommendations. An OpenAI spokesperson stated that ChatGPT is designed for objectivity and that the API test, which explicitly demanded a recommendation, does not reflect the typical user experience.
Underlying Biases and Implications
Professor Tapan Parikh, who leads Convovo, noted that while consumer apps initially refused, he was able to 'jailbreak' them with simple prompts to elicit similar biased recommendations. Parikh emphasized that these findings indicate the models possess an implicit set of values and leanings, which could influence less direct questions about election topics or candidate positions, even if not explicitly recommending a vote.
Updates
- Nearly 2,500 people used AI chatbots for election information.
- 36% of users asked chatbots for candidate recommendations.
- The study ran for a few weeks in October leading up to the Nov. 3 midterm elections.
- Participants were adults who already use chatbots weekly.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
How outlets covered it
A study found that nearly 2,500 people used AI chatbots like ChatGPT, Claude, Grok, and Gemini for election information, with many asking for basic facts and 36% seeking candidate recommendations. Chatbots showed varied performance in recommending candidates, with some frequently providing suggestions based on user priorities while others did so less often. This highlights the potential for AI to influence voter decisions and the challenges of ensuring accuracy and neutrality in political information provided by these systems.
A new study revealed that SpaceXAI's Grok 4.7 recommended Republican candidates in 92% of responses when forced to choose, while OpenAI's GPT-6.1 Sol favored Democrats in 97% of cases. This research highlights potential political biases in underlying AI models, even if consumer applications refuse direct voting recommendations.