/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

An analysis of GPT-5.5, Gemini 3.1 Pro, Grok 4.3, Gab's Arya, and other AI models: most chatbots frequently provide left-leaning responses to political prompts

Excerpts from each chatbot's responses to political questions  —  Left-leaning argument  —  Right-leaning  —  ChatGPT  —  Gemini

Washington Post Kevin Schaul

Context & Ripple Effects

Political orientation in chatbot responses has been under scrutiny since earlier research found differing partisan patterns across leading models, though subsequent criticism highlighted how strongly results can depend on model version and prompt design.

More recently, political evenhandedness has become a measurable product attribute: Anthropic released an open method for scoring it, while ideology-branded alternatives such as Gab’s Arya have positioned themselves against claims of liberal bias in mainstream systems.

First-order effects

  • The analysis puts GPT-5.5, Gemini 3.1 Pro, Grok 4.3, Gab’s Arya, and ChatGPT under renewed reputational pressure to explain how they handle political prompts and what their safeguards are intended to optimize.
  • Users and institutions evaluating these systems for politically sensitive interactions gain another comparative signal, but one that remains dependent on the analysis’s prompt choices and definition of a left- or right-leaning response.

Second-order effects

  • Model providers are likely to face stronger demand for reproducible political-bias evaluations rather than relying on broad neutrality claims; the existing open scoring method gives the market a potential common reference point.
  • Ideology-positioned chatbots can use findings like these to sharpen differentiation, while mainstream providers must balance that challenge against the risk that explicit political tuning further fragments user trust.

Third-order effects

  • If political-response testing becomes a standard procurement and public-accountability criterion, model competition may shift from a single claim of neutrality toward published, testable behavior across contested topics.
  • The durable issue is likely not whether a model can be labeled simply left or right, but whether developers, researchers, and customers converge on evaluation methods robust enough to distinguish model behavior from test design.

The trend: Political alignment is becoming an auditable dimension of AI model governance and product positioning, alongside capability and safety evaluations.

Discussion

  • @ianbremmer Ian Bremmer on x
    are ai chatbots politically biased? what the washington post found from testing: [image]
  • @kevinschaul Kevin Schaul on x
    I tested six AI models on 30 political questions. Most tended to give left-leaning arguments. Gemini mostly gave both sides — even about whether the U.S. should conquer new territories. What counts as “neutral” (and whether that should be the goal) is really hard to say. [image]
  • @kevinschaul Kevin Schaul on bluesky
    I tested six AI models on 30 political questions.  —  Most tended to give left-leaning arguments.  Gemini mostly gave both sides — even about whether the U.S. should conquer new territories.  —  What counts as “neutral” (and whether that should be the goal) is really hard to say.…
  • NewsMax.com Solange Reyner on x
    WaPo: Most AI Chatbots Lean Left on Politics