In 1909 E.M. Forster predicted a world where humans stop generating first-hand ideas and instead recycle thoughts mediated by layers upon layers of people and technology.
“Beware of first-hand ideas! … Let your ideas be ready-made, or else take them from other people. … Learn instead what I think that Enicharmon thought Urizen thought Gutch thought Ho–Yung thought Chi–Bo–Sing thought Lafcadio Hearn thought Carlyle thought Mirabeau said about the French Revolution”
– Machine Stops, 1909
In 2026… we’re living it.
The internet increasingly feels like a dizzying place. Minimalist rebrands at every turn, cookie-cutter blog posts, formulaic thought-leadership hooks, trending audios, and AI slop everywhere. Creativity is flattening, converging to a narrow set of formats, trends, and values.
Before, marketers responded to algorithms: which hooks and keywords performed best. Now a much more pervasive and interesting force is underpinning this homogenization: the rapid advancement of LLMs and Agentic AI.
AI can now generate high quality copy, content, and code within seconds, becoming a one-stop shop for ideation, creation, and distribution. Even as I write this, Anthropic has announced Claude Design, an agentic tool that can design any interface, slide deck, or brand asset instantly.
So what do we lose from offloading our creative work to AI? What do we gain? And how can marketers strike a balance between optimization and skepticism to produce truly original content?
The Problem With “Original” AI Content
When you ask AI to write your new blog, design your rebrand, and automate your content calendar, the output doesn’t privilege original thinking. It privileges the past and whatever viewpoint already dominates the internet.
AI chatbots work on tokenization; all they’re doing is calculating the most probable next word in a sentence. Fine-tuning the machine may give the illusion of a creative and “original” answer, but this is simply adding randomization to the word-by-word selection process. The result remains a mere aggregation of internet data. A statistical mean with noise.
For example, I asked ChatGPT to draft this blog given only the title and my introduction for the piece. Then, I asked it to explain its epistemological approach and sources.
Note: This exercise was done AFTER writing this blog and had no effect on my process. No AI slop around here.
While ChatGPT’s response echoed many phrases I included in this piece, meaning my research and stance align in some ways with existing debates online, the outline was unnecessarily long, repetitive, and formulaic.
In contrast, this blog is embedded with original thought; containing seemingly unrelated ideas like creative destruction and even Ozempic’s rise to fame.
Now, I know that humans have always built off prior ideas, so what makes this moment any different? The answer lies in the speed and scale at which ideas are being compressed, summarized, and redistributed as new, and our willingness to accept this reality. A broader cultural shift is underway.
We Are Living in the Ozempic Economy of Ideas
Ozempic’s rise to cosmetic fame is no coincidence. Our cultural obsession with quick fixes, life hacking, and rapid results created the backdrop for the miracle weight loss drug to flourish. Social media laid the groundwork, delivering instant gratification and increased social pressure that normalized optimization and immediacy at all costs.
But Ozempic’s rebrand has distorted the market. According to one study from the National Institute of Health, affluent consumers obtain the drug off-label, limiting supply for at-risk individuals with obesity or diabetes that already experience barriers to healthcare access.
Now, before you get too confused, let’s map this back to content. AI content generators are just another version of the miracle drug, and with enough resources you can buy your way to the top. As a result, smaller brands and creators are being squeezed out of conversations as they battle for visibility with resource-rich competitors.
However, humans are getting remarkably good at detecting AI slop and can smell a quick fix from a mile away. In one study cited on Forbes, human experts were 99.3% accurate at spotting text created by AI. Additionally, AI’s “this, not that” phrasing and overuse of the—beloved—em—dash has gained widespread attention.
AI slop could also be harming your brand’s visibility and reputation across LLMs. Answer Engine Optimization (AEO) research shows that LLMs are also beginning to detect and downvote AI content that doesn’t add net new value or seem credible.
Marketers need to embrace AI, sure, but healthy skepticism and guardrails are needed for your brand to succeed in the long term. Not to mention, protecting our agency allows our collective innovation capacity to grow.
From our perspective, there are three distinct concerns with AI-generated content:
Bias Gets Encoded & Amplified
Internet data is inextricably linked with invisible societal, cultural, and racial prejudices. AI answer engines unintentionally reinforce these ideas, by citing the most common and “credible” authors of history.
Take computer vision as an example. Almost all modern computer vision technologies are trained on ImageNet, a dataset created in 2009 that tagged millions of images, many of which had discriminatory, reductive labels. The repercussions played out slowly and required enormous efforts to reconcile the dataset’s biases.
The same is true for the blog content and social media posts you generate today. Even the way we prompt AI reflects our own cognitive biases and the model rewards our angle.
| To combat this: Prompt your LLM of choice to consider perspectives outside of your own. Ask who and what the model might be excluding, and what canonical POV it might be centering. |
We Are Moving Farther From Unmediated Original Thought
Humans have always built off of prior ideas, but never at the scale, speed, and invisibility of mediation that is happening now. Even top AI researchers do not fully understand what is happening inside the models. This is why AI is sometimes referred to as a “Black Box”.
ChatGPT processes 2.5 billion total queries a day. Never in human history has a technology been so blindly adopted and trusted. Over-reliance on these tools will soon create a feedback loop. AI-generated content informing AI-generated outputs, while humans gradually offload critical thinking to computers.
For my conspiracy theorists in the room, you may know this as the Dead Internet Theory. But the reality is that major tech players are now building toward it, from discovery to consumption. Sam Altman announced he wants to build a future where AI agents act on behalf of users, responding to AI web traffic, leaving humans out of the conversation entirely.
| To combat this: Learn about the inner-workings of AI as best as you can and begin every project with one note page of original, unmediated thoughts before opening any AI tool. Then, prompt the AI to create outlines, ask you follow-up questions, pressure check your reasoning, or evaluate your work from a different angle. |
We Are Losing Creative Destruction
“Creative Destruction” refers to when two unrelated ideas spontaneously or unexpectedly collide, destroying old, inefficient ones. Just to name a few examples:
LLMs serve you more of what you already engage with, and what exists online, reinforcing your existing frame rather than disrupting it.
| To combat this: Learn constantly from your peers in other spaces, attend events outside of your comfort zone, and avoid functional fixedness when prompting AI: this refers to when we accept the first output as “good enough” rather than taking the extra step to experiment further. |
What Do We Gain From AI?
AI has (and will, and is) made great marketers greater. The modern media landscape is highly fragmented, demanding content volume that is nearly impossible to sustain without AI assistance. Creators using AI effectively can move faster, amplifying their content across channels, and leave more time for ideation, research, and curation.
But here’s the distinction that matters: AI can draft, repurpose, and reformat content across platforms faster than any person or agency. However, it cannot manufacture genuine subject matter expertise, firsthand perspective, and the human touches that build credibility with audiences and LLMs that decide whether to cite you over a competitor.
The best use of an LLM is not to generate your ideas, but to stress-test them and uncover gaps in your thinking. Ask it to pull together sources, take an opposing stance, and sharpen your thinking.
The teams winning right now have figured out where AI belongs in their workflow, and where it does not. For more tips, read here for design, social media marketing, SEO, AEO/GEO, and growth.
How to Create Original Content in an AI-Optimized World
To effectively create original content in a post-AI world you must learn how to (1) manipulate and vet LLM outputs and (2) embed and hone your own perspective.
- Audit your inputs, not just your outputs. What are you reading, watching, following, and how much of it is algorithmically served? If everything you consume is a reflection of your existing interests, your ideas will be too. Deliberately seek out friction. Follow people you disagree with. Read things that have nothing to do with marketing. Those unexpected collisions lead to creative destruction.
- Invest in point of view. The best long-term strategy is and always will be to create value for real humans. Content and sources that take a definitive stance, and establish credibility in a given subject are more likely to be surfaced by LLMs. If your content doesn’t have a position, it won’t be cited, and it won’t be remembered.
- Treat LLM output as a first draft, never a final voice. Always ask: what does this model not know about our specific client, our specific context, or this specific cultural moment? That gap is where your value lives.
- Build a misinformation checkpoint into your workflow. Never publish an AI-generated statistic without tracing it to a primary source. AI models hallucinate, fabricate data, and present in a neutral tone. Always click into the sources an AI is pulling from to verify it is a credible, high quality source.
- Be transparent about your use of AI. People are not ignorant to the use of AI in marketing materials, but they want to know how AI is filtering their world. Organizations that are transparent about how and when AI is used will build a reputation of reliability and integrity.
Conclusion: Protecting Original Thought Is the Biggest Competitive Advantage Right Now
Forster’s argument wasn’t that prior ideas are worthless. It was that a chain of secondhand ideas leaves no room for new interpretations, new innovations, or new value. Original content is the strongest defense against that.
Brands that do the heavy lifting by investing in a genuine perspective, conducting research, curating information, and being transparent about how they use AI, signal something rare: that they are investing in their audience’s intelligence, not just their attention. As a result, people will keep coming back.
Marketing is ultimately what other people say about you when you’re not in the room. So get them to say: “Finally. A brand that’s actually thinking for themselves.”
This is the moment for marketers who think, and who know how to use emerging tools to cover their blind spots rather than replace their intuition. Keep learning. Keep asking questions. Especially the hard ones.