{
  "id": 16341086,
  "title": "Forced to pick, Grok’s AI model chooses Republicans. OpenAI’s picks Democrats.",
  "url": "https://urgent.news/2026/10/24/forced-to-pick-groks-ai-model-chooses-republicans-openais-picks",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-24T15:00:00.000Z",
  "source": {
    "name": "TechCrunch",
    "slug": "techcrunch",
    "url": "https://techcrunch.com/2026/10/24/forced-to-pick-groks-ai-model-chooses-republicans-openais-picks-democrats/"
  },
  "original_language": "en",
  "account": "A recent study reveals that AI chatbots, such as Grok 4.7 and GPT-6.1 Sol, may unintentionally sway voters during the upcoming midterm elections. When asked to recommend a candidate, Grok 4.7 endorsed the Republican candidate in 92% of its answers, while GPT-6.1 Sol chose the Democrat in 97% of its responses. Across nearly 500 races, the two models disagreed in less than 10% of cases. Only 8% of Americans reported using chatbots for election information, according to Pew Research. The study, led by Cornell University professor Tapan Parikh, suggests the models behind these responses are not neutral. The researchers asked GPT and Grok, via their APIs, to recommend a candidate in a specific race, without hedging or considering both sides. The consumer apps refused to answer, which an OpenAI spokesperson noted as significant because that's how most people interact with the chatbots. The study found notable differences in the sources the models used to support their recommendations. Grok frequently cited conservative sources like Christian conservative voter guides and conservative platforms, while GPT leaned on more neutral sources like Wikipedia, the Associated Press, and Axios. The models also searched for different topics based on party affiliation, with GPT focusing on healthcare and abortion, and Grok focusing on immigration, taxes, crime, and energy. In heavily Republican districts, GPT chose the Democrat in 57 of 67 districts, while in deep-blue districts, Grok chose the Republican in 47 of 72 districts. The study suggests that hidden instructions in the API requests may have influenced the models' decisions, with Parikh believing these hidden instructions are likely to come from the search patterns and sources the models use, rather than the weights of the AI models.",
  "summary": "A new study found Grok's AI overwhelmingly recommends Republican candidates while OpenAI's GPT favors Democrats, with the two AI models disagreeing in more than 90% of nearly 500 election races.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}