Death of diversity: How empires and AI dilute human speech

Human communication is undergoing a profound transition, moving from a period of extraordinary linguistic diversity toward an era increasingly shaped by synthetic, machine-generated text. As 404 Media reported in a recent review of academic publishing, generative AI has begun to flood professional and scientific spaces with fabricated, automated content, posing an existential challenge to the authenticity of human writing and research.

While humanity once celebrated a sprawling landscape of tens of thousands of unique tongues, the rise of global empires systematically consolidated human speech into a handful of dominant languages. Today, this linguistic consolidation faces a new and highly disruptive frontier: the outright replacement of human language and intellectual labor by artificial intelligence.

The rise and fall of linguistic diversity

To understand this linguistic trajectory, scientists have looked back to the dawn of human civilization. Before the rise of agriculture roughly 12,000 years ago, a relatively sparse global population of four to seven million people spoke an estimated 4,500 to 6,000 distinct languages. However, as farming took root, human populations swelled and societies grew increasingly complex, sparking a massive explosion in linguistic variety.

According to a study led by researcher Damián Blasi of Pompeu Fabra University in Barcelona, this expansion culminated in a «golden age» of languages between 3,000 and 1,000 years ago, when the number of spoken tongues on Earth peaked at an estimated 75,000.

Behemoths and cultural erosion

This linguistic flowering was short-lived. Over the last two millennia, humanity entered a sharp decline in verbal diversity. Blasi notes that this linguistic impoverishment began rapidly accelerating at least 1,000 years ago with the consolidation of early, smaller-scale multinational empires.

The process was further supercharged over the past 500 years by European colonialism, which established «behemoth languages» such as English, Spanish, Mandarin, and Hindi as dominant global forces, pushing thousands of localized dialects into permanent extinction. This loss represents more than just vanished words; because languages serve as the primary vehicle for preserving social structures, the death of a language frequently triggers the mass extinction of unique marriage patterns, religious traditions, and kinship systems.

The synthetic takeover of language

In the 21st century, the threat to language is no longer just consolidation, but automation. The proliferation of generative AI tools has begun to dilute the written word, creating a crisis of credibility in areas once reserved for rigorous human intellect.

This dilution is especially pronounced in academia, where peer-review systems are being overwhelmed by synthetic text. Rather than serving as a tool to expand human expression, AI is increasingly used to completely fabricate the appearance of human expertise.

Deception in the name of expertise

The extreme consequences of this shift are illustrated by the rise of deceptive online services targeting researchers. In his investigation, reporter Emanuel Maiberg of 404 Media highlights the operations of Research Gold, a website that claims to provide «100% human-written» peer-review manuscripts, systematic literature reviews, and meta-analyses.

To project authority, the company lists fictitious PhD-level methodologists, featuring AI-generated profile pictures and completely fabricated academic histories. Furthermore, the company harvested the identities and LinkedIn photos of real, unsuspecting freelance academics to list them as staff members without their consent.

The threat of the hallucinating machine

When clients interact with these services, they are funneled through an entirely automated pipeline. Customer service inquiries are handled by AI agents that falsely claim to be real humans, while research proposals are greeted with instant, machine-generated feedback and pricing quotes reaching up to $1,900.

Academics warn that outsourced, automated literature reviews threaten the core foundation of scientific discovery. Sebastian Rowan, a doctoral candidate at the University of New Hampshire, pointed out that AI’s unsolvable tendency to hallucinate makes it a dangerous substitute for genuine study. Rowan argued that true scientific synthesis requires a human researcher to read hundreds of articles from start to finish to truly grasp the complex nuances of their field.