In the era of big AI, human writers deliver a bigger payoff for brands. Surprised?

• July 8, 2025
• no comments

By Hollie Shaw, Head of Editorial Services and Executive Strategy

—

The boogeyman of modern editorial content writing is AI-generated misinformation — or even worse, downright fabrication. Who can forget the recent debacle at the Chicago Sun-Times, which published a summer reading list filled with a number of book titles that did not exist?

The spectre of unmediated AI content running in a major U.S. newspaper and through mass media at large is a major concern, obviously.

But there’s another compelling reason why companies should use human-written content if they want anyone to read, watch, or consume it: in the vast ocean of online information, there’s ample evidence that AI-written web content, blogs, and branded thought leadership will travel almost immediately to the internet’s Bermuda Triangle when published, unnoticed and essentially undiscoverable.

That’s because human-generated content is an overwhelming indicator of how frequently and how well your brand shows up in standard web searches, and critically, how frequently and how prominently your brand shows up in the AI results of web searches, AI-driven queries and AI content results.

“If you’re a company that wants to raise its profile or a marketer looking to elevate a brand, you need to go beyond table stakes investments in Google ads and search engine optimization,” says Jessica Savage, our CEO at North Strategic. “And don’t even consider severing thought leadership from a human and handing the writing over to AI.”

Instead, our agencies think about where and how C-suite leaders can influence industry — via earned media, owned media, public events and panel appearances — because that’s where AI assistants go to refuel their engines.

When a human holds the proverbial pen, the content will perform better in backlinks, time-on-page, and social shares — key indicators of expertise, relevance and quality. These metrics influence how search engines rank content, and boost visibility and brand awareness as a result.

This all makes sense: human-generated content is the lifeblood of AI content. You can’t have one without the other. And AI’s large language models, capable of churning out a recipe or a haiku or a form letter in less time than it takes to sneeze, couldn’t exist without earned media.

AI writing, earned media and the necessary human touch

Online news from recognized media organizations (written by humans, for humans!) was and still is one of the main foundational elements used to train the LLMs that power generative AI. Media — earned, owned, paid and shared — are the basic dietary staples of LLMs.

Earned media is prioritized by Google and other search engines as a top source of information and a barometer of trustworthiness. Indeed, Google explains this directly in its guidance for web developers trying to improve an organization’s search engine rankings. Search engines rely on the acronym E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness) in ranking search results.

This is no secret to AI evangelists, and it’s a principle inherent to the future development of AI. David Benigson of Signal AI has repeatedly emphasized the necessity of mediating AI with human expertise to safeguard accuracy and critical thinking. “AI can identify patterns, but human expertise provides the necessary context,” he told Forbes. In other words, human intervention is required to assess risk, identify misinformation, and make informed decisions.

This is particularly important given the desire of organizations to embrace AI while avoiding its biggest risks. Have you ever asked Chat GPT or Microsoft Copilot to execute a piece of writing and then requested a comprehensive list of the sources it used? Did it draw a blank? Or did the content displayed on the other side of AI’s cited source links fail to provide you with clear proof points for claims the AI just made in its own written work?

This glaring lack of “explainability” is one of the biggest hurdles organizations encounter when they try to implement AI successfully, according to a recent McKinsey analysis on AI. “LLMs are often black boxes that do not reveal why or how they came to a certain response, nor what data was used to make it,” the report notes. “If AI models cannot provide clear justifications for their responses, recommendations, decisions, or actions … they will not be trusted for critical tasks.”

The ‘silver bullet’: Human-generated content

Google favours original, unique, and well-researched content over repetitive or AI-generated material. Beyond that, there’s ample evidence AI-generated written content can tank a brand’s search engine ranking while human-written content will boost it and have a more sustained impact for the brand over time.

An analysis by Boston-based SEO strategy firm Backlinko of 912 million blog posts determined that longer, high-quality written pieces indicative of human authorship earned 77.2% more backlinks than the shorter, SEO-optimized pieces generated by AI. Critically, a tiny percentage of well-written pieces generated the bulk of social sharing: 1.3% of the articles accounted for 75% of all shares.

Kristan Bauer, a top SEO consultant and business strategist, detailed this discoverability paradox in a recent LinkedIn post. Her client’s business was indexing poorly on Google, she said, despite having a corporate website with optimal SEO and user experience and zero site security or site authority issues. The client followed a series of Bauer’s initial recommendations and to-dos, but after deploying many of them over the course of nine months its poor ranking persisted.

Finally, the client followed a recommendation of Bauer’s that it didn’t heed initially, one that would have no doubt required more of an up-front investment: replace all written AI content on the website with human-written editorial and programmatic content, and take note of any differences in search ranking.

The results were a fast and dramatic reversal. Impressions grew fivefold in the four weeks after the client replaced AI-written content with high-quality custom content written by a human, and continued to grow from there.

The “silver bullet,” Bauer said, was human-generated content.

How we use AI at North

Admittedly, as a person who’s spent the bulk of my adult life writing professionally and the leader of an in-house writing team at North, I have a vested stake in this.

Our team has been testing AI’s writing capabilities (and not for clients’ written materials, for reasons we’ve made abundantly clear) and compared them to our own since ChatGPT made its free public debut in 2022. I routinely use Microsoft CoPilot as a thesaurus. We all know AI is a useful tool for sparking writing ideas, editing for copy errors, syntax and organization, and condensing internal materials. Publicis, our parent company, has embraced AI in a major way.

But if you care at all about brand clout, it’s clear AI should not be the brain behind your written communications, for now and for the foreseeable future. Make sure a human is powering the keyboard.

To learn more about our strategic content writing team, (all human, last time we checked), contact us today.

Contact