What Does Grokipedia Know About Canada?

As you might have heard, Elon Musk just brought version 0.1 of his new Grokipedia encyclopedia online. It’s no secret that the motivation driving the project was serious and widely documented claims about the neutrality of Wikipedia.

The Current State of Wikipedia

At the forefront of criticism of Wikipedia has been serious research by Pirate Wire’s Ashley Rindsberg, academic David Rozato, and a site called Tracing Woodgrains. But the common denominator between them is that a small number of Wikipedia editors has redefined “neutral” and “reliable source” to mean anything that fits a narrow hard-left worldview. And the organization’s decentralized structure has allowed that group to enforce their ideologies across a growing range of topics.

Twenty years ago, the most powerful argument in support of Wikipedia’s open source model was that a large community of volunteer editors would effectively correct its own errors and biases. In many cases, that did work. Prank edits (the online equivalent of graffiti) were often removed in minutes or even seconds. And disputes were sometimes resolved peacefully in a page’s discussion section.

But even in those early days, personal or ideological conflicts were far more difficult to referee. Many years ago I was involved in an unpleasant edit-war on a Wikipedia page. I gave up after a few rounds, but my trust in the value of the dispute process was permanently damaged.

The academic world’s transition away from objectivity to identity-based narratives has made all that a lot worse. Countless Wikipedia pages are now written to conform to standards where the plain facts have limited influence and it’s often recent academic publications that are promoted as authority.

Web visits to Wikipedia’s millions of pages are currently in free fall. Between March, 2022 and March, 2025, traffic to the site dropped by a whopping 23 percent. For comparison, the ChatGPT site attracted a half billion more visits than Wikipedia in March, 2025. Which, as it turns out, is another part of the problem.

To varying degrees, all of the major AI models have been trained on Wikipedia content. So if you’re uncomfortable with existing Wikipedia biases, you won’t be thrilled by the fact that that bias exerts an unknowable influence over the stuff your favorite AI chat model tells you.

That’s the situation Grokipedia seeks to address. Musk has replaced “armies of volunteer editors” with some iteration of his own Grok generative AI model. He acknowledges that this early release will certainly contain problems. But he’s convinced that even with those problems, this new project will still be significantly more reliable and useful than Wikipedia.

Canada: As Grokipedia Sees Us

The first thing I’ve noticed about this early version of Grokipedia is that the writing feels different. It’s a bit more formal and academic-sounding. It also somehow doesn’t have that obvious “AI-spat-me-out” feel.

The first page I looked at was “Dairy and poultry supply management in Canada”. That entry contains nearly 11,000 words, compared with the comparable page in Wikipedia that’s nearly double the length.

Both the Grokipedia and Wikipedia entries start with a helpful overview section and both are intelligently organized with logical headings. But that’s something else you’d expect.

However, what I’m most interested in is how complex and controversial topics are handled. Here’s how ChatGPT (Grok’s primary competitor) compared the two. “File 1” contained the Wikipedia version and “File 2” was Grokipedia’s entry.

As you should (hopefully) be able to see, overall ChatGPT consider’s the Grokipedia version to be more balanced and better at critically “integrating data and policy analysis”. It’s also written in a style that’s more easily understood by a wider range of readers.

---

---

By contrast, the Claude AI analysis was a bit more critical of Grokipedia’s entry, considering it “Notably less balanced overall”. Although it did feel that Wikipedia had a “heavy reliance on opinion pieces and advocacy sources” and that “debate sections sometimes blur into advocacy”. Here’s what Claude wrote:

---

Document 1 (supply-management-1.md)

Source: Wikipedia article on Canada’s supply management

Strengths:

• Comprehensive historical coverage from the 1930s through 2018

• Extensive citations (over 200 references)

• Presents multiple perspectives explicitly, with dedicated sections on “Debate” and “Supply management debate”

• Includes specific quotes and attributions from various stakeholders

• Documents the evolution of policy positions across political parties

• Provides detailed information on lobbying activities and interest groups

Weaknesses/Objectivity Concerns:

• Contains neutrality disputes (explicitly noted in the “Retail price of milk” section)

• Some organizational issues with repetitive content across sections

• Heavy reliance on opinion pieces and advocacy sources alongside academic work

• The debate sections sometimes blur into advocacy rather than neutral description

• Includes detailed criticisms from think tanks without always clearly identifying their ideological leanings

Document 2 (supply-management-2.md)

Source: Appears to be from an AI-generated reference site (notes “Fact-checked by Grok”)

Strengths:

• Well-organized with clear hierarchical structure

• More recent information (covers through 2025)

• Systematic presentation of mechanisms and effects

• Cleaner separation between descriptive content and perspectives

• More consistent citation style and formatting

Weaknesses/Objectivity Concerns:

• Notably less balanced overall—presents criticisms more prominently in the main text

• The opening summary paragraph contains strong normative language about “net social welfare losses” and “regressive effects” before presenting supporting arguments

• The “Economic Effects” section emphasizes negative consumer impacts more heavily than producer benefits

• Less attribution of contested claims to specific sources in the body text

• The AI fact-checking claim raises questions about verification methodology

Key Objectivity Comparison

Document 1 attempts encyclopedic neutrality but struggles with balance in places. It’s more transparent about controversies and explicitly labels disputed sections.

Document 2 presents a more polished structure but embeds more normative economic judgments in ostensibly descriptive sections, particularly regarding efficiency and welfare effects that remain contested among economists.

Recommendation: For balanced research, Document 1 provides better transparency about debates despite its organizational issues, while Document 2 offers clearer structure but requires more critical reading for embedded assumptions.

---

Let’s explore one more topic that has been known to evoke strong (editorial) opinions: Justin Trudeau. In this case, according to Claude AI, Wikipedia clearly won out over Grokipedia’s entry - at least as far as objectivity goes.

---

Document 1 (trudeau1.md) - Wikipedia-style Article

Strengths:

• Comprehensive, well-cited content with extensive inline references

• Neutral, encyclopedic tone throughout

• Balanced coverage of controversies and achievements

• Detailed chronological structure

• Multiple perspectives presented on controversial issues

• Extensive documentation of sources

Weaknesses:

• Some sections are extremely dense with detail that may overwhelm readers

• Occasional redundancy in covering the same events multiple times

Objectivity: High - Maintains neutral point of view, presents facts with proper attribution, and includes multiple perspectives on contentious issues.

Document 2 (trudeau2.md) - Structured Overview

Strengths:

• Clear, organized topical structure

• Comprehensive coverage of major policy areas

• Detailed economic analysis with specific data points

Weaknesses:

• Significant objectivity concerns:

  • Opening paragraph disproportionately emphasizes negative outcomes (debt doubling, blackface, Emergencies Act)

  • Economic sections heavily weighted toward criticism (”predominantly negative,” “failures dominate critiques”)

  • Selective framing in assessments section: “54 percent viewing legacy as more failures” without comparable positive polling

  • Language choices reveal bias: “modest reductions,” “limited progress,” “stalled,” “superficial”

  • Final “Assessments” section explicitly states negative consensus as fact rather than one perspective

  • Cultural criticism section presents conservative viewpoints as definitive (”performative progressivism,” “elite cosmopolitanism”)

Objectivity: Moderate to Low - While factually detailed, the document exhibits clear editorial bias through selective emphasis, framing choices, and presenting contested assessments as established fact.

Overall Comparison

Document 1 is significantly more objective and follows encyclopedic standards, while Document 2, despite its comprehensive data, functions more as a critical assessment that prioritizes negative interpretations of Trudeau’s tenure.

---

This is how Grokipedia looks just days after it’s first (version 0.1) release and it’s not bad. There may well be bias - and that bias may well extend to Canadian topics - but I suspect that may still be less due to malicious intent than to weak filters.

Insofar as I use online encyclopedias at all these days, at this point I’ll probably go with Grokipedia. Although I’ll definitely keep my eyes wide open for its own peculiar quirks.

---

You might also like: