How AI Affects Content Originality

Two new studies reveal that AI has a homogenizing effect not only on content style but also on content substance. The good news is we're seeing this impact early enough to do something about it. A study highlighted by USC Dornsife suggests AI may homogenize how we communicate and reason, often reflecting a narrow (WHELM) slice of perspective. At the same time, research from Google DeepMind and a coalition of researchers from west coast universities finds that content generated by LLMs can systematically influence how ideas are expressed, making content more similar in both style and substance. So, now we can make even more informed choices about using AI. Yes, AI can streamline content creation, improve clarity and consistency, and scale production in ways we couldn’t before. But it can also nudge us toward sameness in our content and in our thinking. There are many implications here, so I'll focus on some business ones. Let's take high-stake areas like health and finance: > Content may overlook important population differences or edge cases. > Bias in recommendations can reinforce disparities. > “Averaged” thinking can lead to poor decisions when nuance matters most. Let's also think about marketing, where brand voice and messaging are differentiators. If everyone uses the big LLMs the same way then > Messaging starts to sound interchangeable. > Positioning loses its edge. > Brands become harder to distinguish. Again, that's just a sample of the implications. I don't see these research findings as a reason to stop using AI, but they're yet another reason to get smarter about it, especially at scale. We have options. Let's use them wisely. More about the USC study: https://lnkd.in/ef8DVrJw More about the Google DeepMind study: https://lnkd.in/eDQtzY-h More of my perspective on AI options: https://lnkd.in/e7xtWegs #ai #contentstrategy #marketing #health #finance #digitaltransformation #writing #aistrategy #modernization

Everyone fears AI replacing humans. Few notice it’s already replacing originality. That risk has a name: homogenization. Here's the problem with AI training: We're feeding it standardized content. Then we're using it to develop our unique ideas. What comes out the other side? More standardized content. Think about what happens when you put a rough idea into AI: • You start with something raw and different • The AI smooths it out based on patterns it's seen • It restructures your thinking to match what's "normal" • Your unique perspective gets filtered through a sameness engine The output looks polished. Professional. Safe. But the edge? The thing that made your idea yours? It's gone. This is the real risk of the AI conveyor belt. We're not just speeding up production. We're losing originality before ideas ever see the light of day. The organizations I work with that maintain their edge don't abandon AI, but they use it differently. They treat first drafts as first drafts. They protect the messy thinking phase. They make space for ideas that don't fit the pattern. Because here's what I've learned: Great ideas don't come from optimization. They come from the parts that don't quite fit. The parts AI wants to smooth away. ——— ✦ ——— 🌱 More on AI + Workforce Development → Janet Perez

Congratulations: Now You You Can Blend In And Hide Generative AI Is Killing Our Originality And Making Us All The Same If you have always been afraid to stand out have I got the perfect technology for you! Generative AI is sold as a creative revolution. And it gives that illusion. But data shows Gen AI’s outputs are statistically derivative and regressing to the mean. AI’s Outputs Are Predictably Average A 2024 MIT Media Lab study analyzed 10,000 AI-generated artworks from models like DALL·E 3 and Stable Diffusion. The verdict? 85% of outputs were traceable to patterns in training data, mimicking styles or themes from existing works[1].   Text-based AI fares no better. A 2025 University of Chicago study tested 1,000 AI-generated stories and essays (using GPT-4o and Claude 3.5). They scored 30% lower on originality metrics than human-written works, with 70% relying on recycled tropes or structures[2]. The Math of Mediocrity Gen AI’s architecture, transformers trained on web-scale data, prioritizes probability over invention. A 2024 Nature paper explains that models like LLaMA generate outputs based on “maximum likelihood,” favoring common patterns over outliers[3]. In practice, this means AI art leans toward generic aesthetics (think “cyberpunk cityscape”), and AI text gravitates to clichés (e.g., “in a world where…”).  Consider the numbers: a 2025 Stanford analysis found that 65% of AI-generated images on platforms like Midjourney shared visual motifs (e.g., smooth gradients, symmetrical faces) due to dataset biases[4]. Similarly, 80% of AI-written news articles in a 2024 Poynter Institute study used formulaic structures, reducing stylistic diversity[5]. A Creative Drought If AI floods industries with derivative content, what happens to human originality? A 2025 Oxford study of 500 creative professionals found that 60% felt pressured to match AI’s “polished but generic” outputs, stifling their unique voices[6]. Markets are already saturated: 40% of online content in 2024 was AI-generated, per a WebID study, crowding out novel ideas[7]. Gen AI isn’t doomed to mediocrity We are ******************************************************************************** The trick with technology is to avoid spreading darkness at the speed of light Stephen Klein is Founder and CEO of Curiouser.AI, the only Generative AI platform and strategic advisory designed to elevate an individual's and organizational's intelligence. He also teaches AI Ethics at UC Berkeley. To learn more visit curiouser.ai or connect on Hubble https://lnkd.in/gphSPv_e Footnotes: [1] MIT Media Lab, “AI Art Analysis,” 2024. [2] University of Chicago, “Originality in AI Text,” 2025. [3] Nature, “Transformer Models and Creativity,” 2024. [4] Stanford AI Lab, “Visual Bias in AI Images,” 2025. [5] Poynter Institute, “AI in Journalism,” 2024. [6] University of Oxford, “AI’s Impact on Creatives,” 2025. [7] WebID, “AI Content Prevalence,” 2024.

AI was supposed to revolutionize the way we write. But after two years of working with these tools, I'm seeing something that's getting on my nerves. It's not that AI writing is awful. It's that it quietly pushes us all toward safer, more average writing. At heart, AI tools aren’t creative. They’re predictive. They work by analyzing massive datasets to find the most common patterns in human writing. And that means they’re literally programmed to surface boring, average writing. Like: “In today’s ever-evolving digital age…” It’s crystal-clear when you reprompt an AI. AI isn’t particularly good at thinking of unique ideas – so you’ll often get similar sentence structures and phrasing, even when you explicitly ask it for different variations. (And to be clear, you CAN get around this – but it takes thought, effort, and creative prompting.) I once asked an AI founder what he thought AI was going to do, and his response was: “I think AI is going to take over a ton of mediocre writing that’s being done by humans right now.” And that tracks. AI has absolutely nailed a very generic B2B voice. One that’s professional, relatable, and totally ignorable – and doesn’t deliver any original insight or memorable takeaways. It plays into our natural human tendency to avoid risk. We all seek safety and avoid straying from the group – especially when it comes to putting our ideas out into the world. Nobody wants to be wrong. Nobody wants to look stupid. Nobody wants to say something too risky. That’s the real appeal of AI – it can take the scary parts out the content creation process so all we have to do is hit the easy button. Layer that on top of the baked-in tendency of LLMs to pull us toward the average, and it’s a recipe for more mediocre content than ever. So what are we supposed to do? → Identify the places where an AI can be legitimately helpful – and where it’s just getting in the way → Give the AI tons of data on how your audience actually talks and thinks – like customer interviews, online reviews, and conversations with your team → Use AI to reverse engineer the writing you really admire and analyze why it works – so you can break out of bland, homogenized AI writing → Treat prompting like a creative exercise. Try weird shit. Break away from standard templated prompts and try talking to AI like a real person AI isn’t going anywhere. And if we want more than just average writing, it’s on us to find ways to use it creatively.

The AI Slop Flywheel: Every time you engage with something original online, the platform shows you 5 AI copies of the same original thinking. Here’s why there is so much AI and so little human online. AI is good at writing content and creating images that are similar to what is popular now. Most content recommendation systems are built to surface content that is similar to what is popular and what individuals engage with. That incentivizes creators to copy viral content instead of creating original thinking. Most original thought pieces do not get picked up by the algorithm, so they are high effort and low reward. The safe path to influencer status is to monitor social media for an idea or story going viral and use AI to create a copy. That leads to AI copies ruling the platform, forcing out the original thinking it needs to survive. AI copies become the only source of viral content, so the platform becomes a circular echo chamber of low-quality ideas. People looking for fresh ideas and original thinking (the highest value users) leave the platform. Some AI flywheels are ecosystem builders and enablers. Other AI flywheels are destroyers. This is a fundamental platform design tenet that everyone from AI architects to product managers must learn. We must build our platform’s models to amplify the right flywheels and disrupt the destructive ones. This is not just a social media problem. AI flywheels can make or break platforms from e-commerce to finance, marketing, HR, and supply chains. Just look at what AI has done to the job application process, product marketplaces, and ad marketplaces. When I help clients hire AI talent, I am amazed by how few candidates understand positive and negative AI flywheels. We are 4 years into the AI cycle, but most have not upskilled. Those who haven’t are in for an unfortunate surprise when they interview for a new role or apply for a promotion. AI and agentic architecture, platform design, and monetization all follow novel patterns and practices. When the platform’s mechanics reward AI, you get a lot more AI. Will that make or break your platform? Those who understand AI flywheels create amplification and enablement cycles that generate significant ecosystem growth. It takes a few seconds to point out the companies that do not have the right people because their platforms are being disrupted by AI flywheels.

Most people won’t say this, but AI is making content more skippable. In a world where everyone can write faster, prettier, and in perfect structure, the thing that will stand out even more is this: a real thought. Lately, I’ve noticed this in my own behaviour. If the first 2–3 lines of a post feel AI-written, I skip. Not because AI is bad. But because if I wanted a clean summary, I would just ask AI directly. The reason I read someone’s content is different. I am there for their lived experience. Their judgment. Their stories. Their way of seeing the world. The little imperfections that make the thought feel human. That is the part I don’t want polished out. AI will give all of us enormous power. It will help us think faster, package better, and get more done in minutes than we earlier did in hours. Use it. Absolutely. But don’t lose the one thing that actually makes people stop for you. Your originality. Your authenticity. Your identity. Because in this new era, information will be abundant. Original thought will be scarce. And scarcity is what people value. So yes, use AI to speed things up. Just don’t let it flatten your voice. Because the more the internet starts sounding the same, the more people will look for someone who still sounds real. #ai #brand #marketing #content

Creativity is up. Diversity is down. Everyone’s talking about how GenAI speeds up creativity. But speed isn’t the real shift. What’s actually changed is how we create, and who does what. A few years ago, being creative meant starting from zero. Now it means knowing what to keep, what to cut, and what to remix. The best creatives I know aren’t just creators. They’re: ➡️ Curators, picking from AI outputs ➡️ Editors, tightening what matters ➡️ Prompt writers, shaping the next round GenAI doesn’t replace the process. It reshapes it. Taste matters more than originality. Workflow matters more than perfection. But here’s the catch: When everyone uses the same tools in the same way, The work starts to look the same. Sameness scales. A Science Advances study showed that while GenAI boosts individual creativity, it reduces the diversity of what gets made. Polished? Yes. Different? Not always. If we’re not careful, AI might make us more productive but less original. So the next creative skill isn’t generation. It’s direction. → Inject human unpredictability → Break the template → Ask better questions than the model can answer Curious: What’s one way you’re protecting originality in your creative or product work? Let’s trade notes 👇 🖼️ Image: AI-generated via Ideogram

The biggest threat to content creators isn't AI replacing us, it's creators using AI as a crutch and losing what makes them unique in the process. I've been watching creators rely so heavily on copy-and-paste AI outputs that they're forgetting to develop their own voice, perspective, and authentic point of view. While AI tools are incredible for efficiency and ideation, your unique voice and perspective are what will make you stand out as these tools become more accessible to everyone. The creators who thrive long-term won't be the ones with the best AI prompts; they'll be the ones who use AI to amplify their authentic voice, not replace it. So when Jason Harvey asked me a question about how I feel about AI at BET 's Experience, I had to answer honestly. AI still needs an input, and that input should be "you". Your experiences, your perspective, your voice. As AI gets more advanced and eventually requires just "a press of a button," the only thing that will differentiate you is what you bring to the table that's uniquely yours. My advice? Use AI as your assistant, not your replacement. What's one thing about your voice or perspective that AI could never replicate? That's what you need to lean into more. #ContentCreation #AIStrategy #ContentBranding

The biggest risk with AI is that everything starts to look and sound the same. I’ve been having this conversation a lot lately, whether with colleagues, friends, or on stage. “How can we use AI without losing originality?” Here’s the hard truth: 👉 AI does not create original thought. 👉 It can’t generate emotion or conviction. 👉 And it definitely can’t replace your lived experience. Even if you use five different AI models to create content, be it images, videos, or posts, they often end up looking very similar. But give that same brief to five different team members and you’ll get very different treatments. That’s the difference. Humans(we) bring heart, context, and perspective. And yet, for a while now, a lot of content is feeling the same - overtly polished, optimised, and lacking depth. Why? Because our own voice is missing. 💡Here is a 3P Framework that can help you check and dial into more Authentic content: 1️⃣ Perspective: Your unique lens, shaped by your experiences and opinions. 2️⃣ Personalisation: Your story, your quirks, your voice. 3️⃣ Purpose: The “why” behind your content, the mission that drives it. Check your work against the 3Ps to ensure that it is distinct and differentiated. If you’re working in marketing, building your career, or scaling a business, and using AI, remember this: "AI is your assistant, not your author". Authenticity isn’t optional anymore. It’s your differentiator. #AI #Marketing #Content #LinkedInNewsEurope

Everyone in marketing is talking about AI. How it makes us faster. How it scales our output. How it fills content calendars in seconds. But here’s something I need you to pause and reflect on: 🧠 Is it making us think less? MIT Media Lab just ran an EEG-based study to find out what’s happening inside our brains when we use tools like ChatGPT to write. The results? 👇 → Lowest brain activity: The ChatGPT group showed the weakest neural engagement - even lower than those using Google Search. → Weaker memory & creativity: They struggled with recall and often defaulted to copy-pasting, leading to bland, formulaic content. → Little to no ownership: Participants said they didn’t feel like the work was truly theirs. Meanwhile, the “brain-only” group - no tools, just their thoughts - had the strongest neural activity, the most original essays, and felt the most connected to what they created. Now let’s bring this back to us - the marketers, strategists, creators, copywriters. The ones shaping brand voices. Building authority. Creating thought leadership. 🚨 If we all default to GPT… → Will everything start sounding the same? → Are we diluting our originality in the name of efficiency? → And what happens to that intangible but real thing called brand voice? AI is a powerful tool. No denying it. But when it stops enabling us and starts replacing our thinking, the quality of what we create and who we are as creators is at stake. So ask yourself this the next time you hit ‘Generate’: "Is this me thinking, or just me outsourcing my thoughts?" #cognitivedebt #brainonllm #yourbrainonchatgpt #marketers #creatives