The Wits Lab: Why AI content saturation is pushing brands back to human writers
Manasi Maheshwari’s July 2026 piece for The Wits Lab makes a data-backed case for what many editors have noticed anecdotally: consumer tolerance for AI-generated content dropped sharply between 2023 and 2026, and the flood of mass-produced AI material has made authentic human writing strategically valuable in a way it was not before the flood.
The numbers. The article draws on several surveys and studies. The most striking: a 2025 Muse Report by Billion Dollar Boy surveyed 6,000 respondents across markets and found consumer preference for AI-generated content dropped from 60% in 2023 to 26% in 2026 — a 34-point decline over three years. UCLA research found that consumers now identify AI-generated text with 76% accuracy, and 83% of those who detect it actively avoid that content. On the performance side, human-written content generates 5.44 times more organic traffic than AI-written equivalents, and brands with distinctive voices see 20% higher customer retention rates.
These numbers do not suggest AI writing has no place — the article does not argue that. They suggest that undifferentiated, generic AI output has stopped performing, and that the gap between “some AI use” and “AI-generated everything” is now visible in traffic and engagement data.
The Coca-Cola case. The article uses Coca-Cola’s December 2025 “Holidays Are Coming” remake as its central example. The company released an AI-generated version of its iconic seasonal ad; consumer response was immediate and negative, with viewers describing it as “soulless” and “empty.” The backlash spread quickly and became a reference point in the industry conversation about where AI-produced content fails. Maheshwari uses it to illustrate that audiences detect automation not just in text but in visual and audio contexts — and that detection erodes trust.
What writers can take from this. The article’s practical advice centers on differentiation rather than abandonment. The recommendation is not to stop using AI tools but to direct them toward research, distribution, scheduling, and repurposing — work where speed and scale matter and where brand voice is not at stake. Reserve human judgment for content that carries the brand: named bylines, thought leadership, customer-facing narratives, and anything that represents the organization’s point of view.
Maheshwari also highlights transparency as a factor. Disclosing AI use where it is present was found to be more trust-preserving than concealing it — readers who know AI was involved and see the evidence of human editing and oversight respond better than readers who discover AI use later.
Who it is useful for. Content managers and editors at brands that have been scaling AI-generated content and are starting to see diminishing returns. Also useful for individual writers who need data to support an argument to clients or leadership that human writing is not just a soft preference but a measurable performance differentiator. The article’s statistics are cited with sources, which makes them usable in internal conversations.