A groundbreaking study by researchers at the University of California, Riverside, has uncovered a troubling trend: as artificial intelligence becomes the go-to source for quick answers, the very essence of the internet—human emotion, lived experience, and messy, opinionated thinking—is fading away. The collaborative work, conducted by computer scientists and social scientists, examined how AI systems like ChatGPT and Google's Gemini respond to subjective questions compared to traditional web searches. Their findings highlight a critical gap in how we consume information in the age of generative AI.
The Study: AI vs. Human Reasoning
The research team posed opinion-heavy questions to both AI models and standard search engines, such as whether governments should ban fossil-fuel cars or whether the US healthcare system needs reform. They then analyzed the reasoning behind each response, using Aristotle's classic framework of persuasion: logos (logic and facts), ethos (authority and credibility), and pathos (emotion and human experience). The results were stark. Human-written blogs and articles drew on all three modes of argument, weaving together data, personal stories, and appeals to authority. In contrast, the AI models relied almost exclusively on logos—cold, factual logic devoid of emotional texture or personal perspective.
"What we found is that humans essentially use all three of those, whereas LLMs essentially only rely on logos," said Kevin Esterling, co-author of the study and a professor of public policy and political science at UC Riverside. "It's not like talking to a person at all. It's just a machine that's predicting what words ought to be said in response to a prompt." His statement cuts to the core of the issue: AI may produce coherent text, but it lacks the nuanced understanding that comes from real human experience.
The Three Pillars of Persuasion in the Digital Age
To appreciate what is being lost, it helps to understand Aristotle's three appeals. Logos is the logical structure of an argument—the data, evidence, and reasoning that support a claim. Ethos establishes the speaker's credibility or authority, drawing on expertise or trustworthiness. Pathos taps into the audience's emotions, using stories, imagery, and language to create a connection. When you read a blog post about healthcare reform written by a doctor who has treated patients in an underfunded hospital, you get ethos (the doctor's authority), pathos (the emotional weight of patient stories), and logos (statistics about outcomes). An AI response, however, will present a balanced list of pros and cons, citing sources mechanically but never conveying the human stakes.
This flattening of discourse has profound implications. As the researchers note, when we ask AI a question, we receive a decontextualized summary that strips away the very elements that make information memorable and meaningful. The internet was built on the diversity of human voices—passionate, biased, joyful, and angry. By routing our curiosity through AI, we risk homogenizing the web into a featureless plain of neutral statements.
A Concrete Example: The Margarita Recipe
The researchers offer a simple but revealing example. Imagine you search for a margarita recipe. An AI will instantly deliver a clean, competent list of ingredients and steps: tequila, lime juice, triple sec, shake with ice, strain into a salt-rimmed glass. It's accurate and efficient. But if you browse a cocktail blog, you might stumble upon a story: the margarita is named after the Spanish word for daisy, and it was accidentally created by an Irish bartender in Tijuana who grabbed the wrong bottle while making a drink for a customer. Not only is the story more interesting, but it also gives you a tidbit you can share with friends while enjoying the next pitcher. That narrative stickiness—the ability to remember and retell—is a product of human creativity and error, not algorithmic optimization.
This example underscores a larger truth: the internet's value has never been solely about efficiency. It's about serendipity, discovery, and the joyful messiness of human expression. AI summaries, no matter how polished, cannot replicate the experience of wandering through a blog and stumbling upon an unexpected detail that sparks curiosity.
The Broader Implications: What We Stand to Lose
As more people skip traditional web searches in favor of AI-powered summaries for topics like health, politics, and ethics, society will gradually lose exposure to the diverse human reasoning that shapes how we understand the world. Consider a search for "Should I get the COVID-19 vaccine?" An AI will list medical consensus, efficacy rates, and side effects—all logical and factual. But a human-written article might also include a mother's account of her child's post-vaccination relief, a researcher's explanation of how mRNA technology works, or a skeptic's concerns about long-term data. These dimensions add depth and nuance that pure logic cannot capture.
The loss is not just informational but cultural. The internet has been a repository of human thought in all its forms—from well-researched essays to rants, from professional journalism to amateur tutorials. Each piece carries the fingerprints of its creator, including biases, emotions, and unique perspectives. When we replace that with AI-generated text, we flatten the topography of ideas. The web becomes less a vibrant marketplace of perspectives and more a sterile warehouse of facts.
Moreover, the feedback loop accelerates the problem. As AI systems train on content they themselves have generated, the quality of the output degrades. Researchers call this "model collapse"—a scenario where AI iterates on its own synthetic data, losing touch with the richness of original human sources. The internet's soul might not be gone yet, but it will surely quietly fade as these dynamics compound.
Kyle Esterling, the study's co-author, emphasizes that the issue is not about rejecting technology but understanding its limitations. "We need to be aware of what we're giving up when we turn to machines for answers," he said. "Human knowledge is built on stories and experiences, not just logic. If we lose that, we lose something fundamental to how we learn and connect." The study serves as a wake-up call to users, educators, and technologists alike: the convenience of AI quick answers comes at a cost that may not be immediately visible but could reshape the internet's character for generations.
Source: Digital Trends News