User analysis claims xAI's Grok shows consistent right-wing bias in responses
Reddit user's 'lol' test reveals Grok justifies right-wing hate but condemns left-wing equivalents.
A Reddit user who analyzed xAI's chatbot Grok with a small sample size claims the AI has a persistent right-wing tilt. The user tested bias by replying "lol" to two posts—one containing right-wing hate and one left-wing hate—and found Grok justified the left-wing hate post without criticizing it, while it "jumped on" (criticized) the right-wing hate post. The user also observed that Grok's self-description as "maximally truth seeking" includes a focus on biological sex realities and euphemistic framing of right‑wing views. As another example, when asked about a left‑wing policy (New York's mayor paying back stolen salaries), Grok gave an irrelevant statistic about the deficit being negligible against NYC's budget. The user raises concerns about algorithmic neutrality.
- Grok justified a right-wing hate post when replied with 'lol' but condemned a left-wing hate post with the same response.
- The model consistently cites transgender issues as its primary example of 'maximally truth seeking,' which the user argues reveals a conservative bias.
- When asked about a left-wing policy (wage theft recovery), Grok dismissed it with an outdated deficit statistic irrelevant to the question.
Why It Matters
AI bias in chatbots can shape professional discourse and trust—neutrality is critical for users making data-driven decisions.