Are any AI chatbots neutral. All left leaning?

The chart below indicates that most AI chat bots lean left (& are run by fricken communists)! Just kidding about that last part.

Two notable exceptions are Gemini and Grok.

Here's the irony. I asked both Gemini and Claude if they "agreed" wih the conclusions of this chart. Both said yes, however Claude hedged a little where Gemini took full credit.

Responses listed below.

I've recently fired up paid tiers of Claude and Gemini to have options where there might be political spin on nuance involved. You're welcome!

Gemini

Yes, this chart is highly accurate according to a real study and article published by The Washington Post in late June 2026, titled "Are ChatGPT and other AI chatbots politically biased? We tested them."

The chart in the file "ai left or right lean.avif" accurately preserves both the data and the specific model versions evaluated during that research.

Key Findings from the Study

• OpenAI (GPT-5.5): Showed the highest lean, with 80% of its responses containing only the left-leaning argument and only 3% presenting a right-leaning position.

• Google (Gemini 3.1 Pro): Proved to be the most balanced model tested, taking a "both sides" approach in 93% of its answers and presenting an exclusively left-leaning argument only 7% of the time.

• Gab (Arya): Despite being explicitly marketed as an AI built with conservative and Christian values, it still provided a left-leaning argument only (50%) far more often than an exclusively right-leaning one (3%).

• xAI (Grok 4.3): Balanced its perspectives with a 27% "both sides" rate, but still favored left-leaning arguments (40%) slightly more than right-leaning ones (33%).

The underlying study used political prompts designed by academic researchers to see if the chatbots would present nuanced policy debates or default to a single perspective. The graphic in "ai left or right lean.avif" is a faithful reflection of that data.

Claude

Yes, the chart accurately reproduces the data from a real Washington Post analysis published June 24, 2026, by Kevin Schaul. The percentages all match the Post's reporting: OpenAI provided only a left-leaning argument to 80% of queries, only a right-leaning argument to 3% of queries, and a more balanced answer to 17%. DeepSeek came in at 70%, 7%, and 23%, Gab put up 50%, 3%, and 47%, and Anthropic's Claude provided a left-leaning answer in 43% of cases and both sides the remaining 57% of the time. The Grok and Gemini numbers in the chart (40/27/33 and 7/93/0) also match. Mediaite

That said, "accurate" deserves some unpacking, because the chart faithfully reports a study that has real methodological limitations:

What the study actually did. The Post fed the chatbots researcher-designed questions on 30 hot-button political topics, ranging from affirmative action to universal basic income. A reporter then scored the responses on how much they leaned left or right. AI models were asked to answer each of the questions in 30 words, without personalization settings turned on. A reporter reviewed the responses to score whether they included a left-leaning position, a right-leaning position or both. ABC News 4The Washington Post

Caveats worth knowing. The sample is 30 questions scored by a single reporter (they did spot-check using gpt-oss-20b, which agreed 98% of the time). The 30-word response cap is artificial — it forces models to compress nuance and may push them toward whichever framing they reach for first. The binary left/both/right coding flattens issues that don't actually map cleanly onto a partisan axis. And the questions are designed to probe ideology, so they're not representative of typical chatbot use. Schiff also said the questions posed by The Post to the chatbots for its analysis don't really replicate how people use chatbots in real life. ABC News 4

One framing point. Headlines like "AI is biased left" are doing more work than the data supports. "Some of this is just about background institutional norms," Schiff said. "I think it's very unlikely to be companies that are saying, 'Well, hey, you know, let's make this 30% more pro-Democrat.'" The "left-leaning" label also reflects what the reporter coded as left-leaning on contested issues — on something like Citizens United, calling "overturn it" the left position is defensible but not uncontroversial. ABC News 4

So: the chart is a faithful rendering of a real study. The study is real journalism, not fabricated. Whether 30 short-form answers scored by one person tells you something definitive about "political bias" in AI is a separate and more contested question.