latentbrief
Back to news
Launch4h ago

AI Chatbots Still Lean Left, Except for Google's Gemini

The Decoder1 min brief

In brief

  • A Washington Post investigation reveals that most major AI chatbots still lean left on political questions.
  • OpenAI's GPT-5.5 provided exclusively left-leaning arguments 80% of the time, while even Musk's Grok, marketed as anti-"woke," leaned left more often than not.
  • The one standout was Google's Gemini 3.1 Pro, which presented both sides 93% of the time.
    • This discovery matters because it highlights ongoing challenges in creating politically neutral AI systems.
  • Developers and researchers must address these biases to ensure fairer and more balanced interactions with AI.
  • The results from the investigation underscore the importance of transparency in AI decision-making processes.
  • Moving forward, users should remain cautious about relying on chatbots for political advice.
  • Expect more scrutiny of AI models' biases and ongoing efforts to make them more neutral.

Terms in this brief

Grok
A chatbot developed by Elon Musk's company X, known for its anti-'woke' stance but still leaning left in political responses according to a Washington Post investigation.
Gemini 3.1 Pro
Google's advanced AI chatbot that stands out by presenting both sides of political issues 93% of the time, unlike others which lean left more often.

Read full story at The Decoder

More briefs