Ahrefs Employs AI Across All Content, Defining and Combating "AI Slop"

Every article published by Ahrefs incorporates artificial intelligence, a practice that the company emphasizes is not about replacing human effort but augmenting it to produce higher-quality content. The core issue, according to Ahrefs, is not the presence of AI but the lack of human oversight, judgment, and original contribution, which they define as "AI slop." This form of content, they argue, shifts the burden of effort from the creator to the reader, who must then discern trustworthiness, relevance, and understanding. Ahrefs outlines a four-pillar approach to leveraging AI responsibly, ensuring that its use enhances, rather than diminishes, the value of their published work.

The first pillar, "Do the human work before you draft," underscores the necessity of significant human intellectual investment prior to AI’s involvement in content generation. Traditionally, writing itself was the most time-consuming aspect for content creators. With AI drastically reducing the effort required for drafting, Ahrefs advocates for redirecting this energy upstream to crucial pre-writing stages. This includes meticulously deciding what topics warrant coverage, how to best structure articles, and identifying unique human contributions. Before an AI is prompted to generate prose, Ahrefs insists on a clear premise that addresses fundamental questions: what is the topic, what is the angle, who is the target audience, and what is the core value proposition? While AI can assist in answering these questions by analyzing search results, challenging angles, or identifying missing evidence, the ultimate decision-making authority must remain with the human.

Crucially, AI needs more than just internet data; it requires "raw material" that it cannot produce independently. This includes firsthand interviews, proprietary internal knowledge, unique datasets, real-world demonstrations, accounts of failed experiments, and specific examples drawn from the company’s own work. Ahrefs highlights the development of internal "Source of Truth" tools, such as Mateusz’s app built in Letaido, which functions as a searchable library of facts, explanations, product details, and how-to guides essential for content creation. On a more personal level, the author describes a shift from typing to dictation using tools like Wispr Flow. This method allows for the unedited, stream-of-consciousness transfer of ideas, including uncertainties, caveats, and nascent connections, which often contain the most valuable substance for AI to then organize and interrogate. The goal is to provide AI with rich, human-generated material rather than asking it to invent substance from scratch.

The second pillar, "Create places to stop and pass judgment," emphasizes the importance of breaking down the content creation process into distinct stages, with human decision points in between. AI’s ability to generate polished prose in a single output can mask the numerous decisions made about research, angle, structure, claims, and tone. To counter this, Ahrefs advocates for a workflow that mirrors human editorial processes, where each stage—research, content gaps, outline, and draft—is saved separately. This allows for inspection, correction of problematic outputs or instructions, and restarting from an acceptable point. The principle can be applied even without complex pipelines by creating distinct phases in the writing process, such as brainstorming, outlining, and drafting, with mandatory human judgment at each juncture. This could involve requesting more research, deleting sections, altering the angle, or even abandoning the article. This structured approach combats the "sunk-cost fallacy" often exacerbated by AI, where the immediate availability of polished text discourages questioning its fundamental existence or merit. The focus shifts from simply increasing word count to making better decisions at each stage, preventing AI from turning creators into editors before they have fully engaged as thinkers.

The third pillar, "Spend the time AI saves on making better content," addresses how efficiency gains from AI should be reinvested. While Ahrefs acknowledges using AI for automation of tedious tasks—like the Data Refresh Hub, which automates monthly updates for 12 datasets, saving a significant amount of manual work—they argue that volume is not the sole or primary return on efficiency. The company’s approach prioritizes investing saved time and resources into enriching content in ways that were previously unfeasible. This includes performing simple data analysis, prototyping free tools, building interactive article elements, improving user interface design, or automating parts of research projects. AI lowers the barriers for content marketers to engage in these more complex tasks, even if the initial results are not perfect and specialist involvement is still required for high-stakes endeavors. This allows for richer content experiences, such as transforming basic data embeds into interactive exploration tools or enabling content marketers to create quizzes, free tools, or data visualizations without lengthy waits for developer resources. The distinction lies in using AI to reduce the work behind each article versus using it to undertake previously unaffordable tasks. The critical question posed is not how many more articles can be published, but what new elements can now be incorporated into an article that were once impossible. The goal is to expand individual writer capabilities without compromising the standards for published content.

The fourth and final pillar, "Give every article an owner and a second human," emphasizes the indispensable role of human accountability. Even with extensive AI involvement, a specific individual must ultimately own the outcome of the content. This owner must possess sufficient topic knowledge to validate claims, correct misinformation, explain the provenance of evidence, and confidently attach their name to the published work. This is why Ahrefs’ Director of Content Marketing, Ryan, reviews every word of every article published on the Ahrefs blog. His industry expertise allows him to challenge premises, question evidence, identify generic sections, and ensure the article truly fulfills its promise to the reader. The presence of a human in the loop is rendered ineffective if that human merely rubber-stamps AI output; they must possess the knowledge, authority, and willingness to reject substandard work. Consequently, Ahrefs maintains that even AI-generated or AI-assisted content requires significant human hours from both the writer and an editor to ensure its publishable quality.

In conclusion, Ahrefs asserts that AI-generated content is not inherently "slop." While their own studies indicate that heavily AI-generated pages can rank, a trend of declining search performance is observed as AI use increases. The likely cause, they posit, is not a Google penalty against AI content but the common practice of companies using AI to circumvent essential human work, leading to the publication of more average content. The fundamental issue, therefore, is not AI itself but the abdication of responsibility. AI makes execution cheaper, which should ideally provide more time for the judgment, evidence gathering, experience sharing, and idea generation that make content valuable. The pivotal question is not the quantity or origin of the words, but whether a human understood, judged, verified, and ultimately stood behind the content that was published. If no human truly owns the result, it is deemed "slop."