Elon Musk, Grokipedia, and the battle for truth

Wikipedia itself began in 2001. Its founders, Jimmy Wales and Larry Sanger, also had a revolutionary spirit and a sense of cultural revolution. The idea was to enable people everywhere to build open-source knowledge together, with freedom and popular creative control. It evolved from an earlier project called Nupedia, which failed because its expert review process was too slow. 

Wikipedia has grown into the largest knowledge bank in history, with over 60 million articles written in 300 languages. It introduced the principle of editorial consensus, whereby volunteers discuss disagreements openly on talk pages, reaching a balanced version that reflects different perspectives. 

This has left it open to accusations of bias. Each language community reflects its own political and media culture. Over time, editorial schools have emerged within Wikipedia that dominate the writing of certain topics.

It was here that Musk saw danger: neutrality turning from a guiding principle into an empty slogan. Along with several conservative voices, he argued that Wikipedia backed a mainstream narrative, particularly on issues of gender identity and politics.

Musk's power grab

Some believe that Grokipedia is an attempt by Musk to concentrate power. Yet criticism of Wikipedia itself, such as accusations of bias, has also come from those closely associated with it, including one of its founders.

In October, Sanger published what he called the 'Nine Theses for Fixing Wikipedia', a tacit admission from one of its founders that it had drifted from the principle of neutrality. 

He called for an end to Wikipedia's anonymity, saying its influential editors should be identified. He criticised the use of 'unreliable' sources, which he said were intended to silence conservative voices.  

Musk quickly endorsed Sanger's remarks, reposting them on X with the comment, "Neutrality can only be restored through transparency." But just days later, xAI announced that its new encyclopaedia was ready to launch.

In the first few days after its release, Grokipedia became a vast experimental space where AI met politics, ideology, and media. Within hours, the number of visitors exceeded 25 million.

Social media platforms were flooded with side-by-side images of articles sharing the same title, such as 'Gaza War 2023'. On Wikipedia, the subtitle read 'The Conflict between Israel and Hamas', while Grokipedia's version was titled 'Israeli Aggression on Gaza'. The difference appeared to be merely linguistic, yet it was enough to ignite a heated discussion about who holds the power to name and frame events.

Supporters said neutrality does not mean the absence of perspective but rather the inclusion of all sides without selective human intervention. Critics disagreed, saying machine learning cannot grasp the moral and cultural depth embedded in human language. For them, bias is not measurable in figures. It depends on an outlook shaped by history and identity.

Amid the debate and the growing rivalry between the two platforms, Grokipedia introduced a new feature called 'Chronological Perspective'. It allows readers to trace how any piece of information has evolved over time, as well as to observe how wording or references have changed through different periods.

Many hailed it as a breakthrough in cognitive transparency—a way for users to witness firsthand how facts themselves shift across the years.

A potential 'cognitive hazard' 

The greatest challenge for Grokipedia is an ethical one. The xAI team acknowledges that Wikipedia was one of many resources used to train the system, but emphasised that "the model does not copy texts; it learns patterns from them." This raised a complex question about intellectual property in the age of AI, the boundaries around which have become blurred.

A new movement began to form among freelance writers and journalists who viewed Grokipedia as a chance to reclaim the individual's voice, in contrast to Wikipedia's anonymous collective editing model. 

Wikipedia introduced an optional feature called 'cognitive signature', which allows contributors to sign articles under their real names or pseudonyms, with links to their previous work. Writing for it shifted from a collective and anonymous effort to a personal practice rooted in digital reputation. Many saw this as a revival of the faded concept of the individual author in an era that had celebrated open collaboration.

The Grokipedia project, in contrast, is the product of Musk's philosophy on the relationship between humans and machines. It has since faced growing criticism from academic and research institutions that view it as an attempt to monopolise the process of truth-making through AI. 

It has been described as a 'potential cognitive hazard', with the warning that manipulating bias-assessment algorithms could reshape history to serve ideological or political agendas. What's more, how can a platform owned by X claim to be a neutral authority when it ultimately answers to Musk himself?

Grokipedia and the three layers of AI

In its early days, Grokipedia relied on three layers of AI. The first analysed the language and extracted concepts, the second compared them with opposing sources to measure bias, and the third produced a final text based on what it called a 'narrative balance of probability'. It constructed paragraphs that accommodated as many viewpoints as possible without clearly favouring one over another. 

Musk described the system in an interview as "an intellectual court that does not issue verdicts but simply presents evidence to the public."

Yet, as the platform gained more users, it became clear that its notion of neutrality did not align with the human understanding of the concept. Some articles felt distant and devoid of emotion, while others were so mechanical that readers began to miss Wikipedia's more human tone, with all its flaws and biases. 

The contradictions within narratives

Nonetheless, Grokipedia has exposed contradictions within historical narratives. For instance, the platform could reveal how Russian and Ukrainian accounts describe the same battle in opposite ways, offering readers a rare glimpse into the relativity of truth itself. It is also a philosophical experiment, an attempt to turn AI into what Musk called a "guardian of meaning".

In his first interviews after the launch, Musk introduced the concept of 'statistical truth', arguing that truth no longer stems from a single authority but from the quantitative aggregation of opposing perspectives. According to this logic, every opinion contains a fragment of truth, and every fact requires enough contrasting viewpoints to settle at an average value that can be considered objective. 

Several intellectuals have observed that Musk's vision is less a cognitive revolution than a digital reformulation of Gustave Le Bon's 19th-century notion of the 'collective mind'. 

Read more: Rapid tech innovation brings AI closer to consciousness

The project itself exposes a philosophical paradox. While it claims to eliminate human bias, it still depends on data, history, and culture, all of which are imbued with that same bias. The machine can only learn from what humans provide, and what humans provide is inevitably shaped by their subjectivities and ideologies.

This paradox has led some AI scholars to ask an unsettling question: can AI generate a new kind of automated bias, one different from but no less dangerous than human bias? 

Musk has tried to address these concerns by insisting that the system is constantly learning and self-correcting. But the fundamental question remains: who teaches the teacher?

The absence of individual responsibility

Intellectually, Grokipedia represents a shift from an 'encyclopaedia of knowledge' to an 'encyclopaedia of consciousness'. It does not simply store information but attempts to understand how beliefs are formed. 

Perhaps the most troubling aspect of Grokipedia's philosophy is the absence of individual responsibility. Wikipedia shows who edited an article, who reviewed it, and who rejected certain changes. In Grokipedia, however, AI makes these decisions based on hidden criteria. 

Neutrality here is not the result of human discussion but of closed calculations that cannot be questioned. As one expert put it: "In Wikipedia you can disagree with a person, but in Grokipedia you disagree with a machine that does not respond."

Musk has often clashed with the West's media. He accuses it of institutional bias against freedom of expression. In Grokipedia, he appears to be trying to create a platform that bypasses journalists, editors, and the traditional ways of marshalling knowledge. 

Through it, Musk is positioning himself as the chief editor of the AI age. Rather than a writer or observer, he is more the architect of a new way of producing knowledge. In this, he is challenging not only Wikipedia but the broader cultural foundation of Western modernity, where, for now, humans, not algorithms, have the final say.

Musk's ambition is a step towards rethinking humanity's relationship with knowledge. It opens the door to a new phase of engineered consciousness. Grokipedia aims to structure meaning and memory.

If this pattern continues, the future may bring a troubling equation in which beliefs are shaped not in universities or newsrooms, but in the research labs of powerful corporations that control algorithms.