There’s a quiet but fairly profound shift happening in the way the internet is being made. Not viewed. Not searched. Not monetised. Made.
For the first two decades of the social web, content still had a basic human bottleneck. Someone had to write the blog, shoot the video, design the image, record the voiceover, edit the clip, translate the campaign, brief the designer, commission the photographer, upload the product copy, crop the assets and hit publish. The internet was already overwhelming, obviously, but it still carried the fingerprints of human effort.
That bottleneck is disappearing.
AI generated content has turned content creation into content production. What once required teams, time and a reasonable amount of creative pain can now be generated, varied, localised, remixed and distributed at a scale that would have sounded absurd only a few years ago. Synthetic images, AI-assisted articles, generated video, cloned voices, virtual influencers, automated ads, machine-written product descriptions and synthetic social posts are now part of the everyday texture of digital life.
The numbers are still difficult to measure with total confidence, because detecting AI-generated material is messy and imperfect, but the direction is obvious. That does not mean the ‘human’ internet is dead in the water though, despite what the more melodramatic corners of the web might suggest. But it does mean the internet is becoming more synthetic. More automated. More probabilistic. More scalable. More polished. And, in many cases, more forgettable. Indeed, how many think pieces have you read already this month on the “deluge of AI slop”?
Simon Manchipp, Founder at SomeOne, puts it with the sort of bluntness that tends to cut through an industry fogged up by softer language:
“The majority are flooding the internet with digital landfill. When everyone can generate a million perfect, soulless images a minute, perfection becomes completely worthless. The new luxury is imperfection. The new premium is a pulse. If it looks hand made, and even better if there's proof it is, you'll make more of an impact.”
That phrase, “digital landfill,” is hard to shake because it captures the central risk of the synthetic content boom. The problem isn’t simply that machines can make content. The problem is that businesses can now make huge amounts of content without necessarily adding much meaning to the world.
And yet, this isn’t a simple anti-AI argument. Synthetic content can be useful. It can make campaigns more accessible, more personalised and more globally relevant. It can reduce production barriers. It can help smaller teams do bigger things. It can allow brands to speak across languages and formats with a speed that would have been impossible in the old model.
Sarina Da Costa Gomez, ECD at Particle6 Productions, is right to resist the lazy assumption that synthetic automatically means bad:
“People will always respond to good, emotionally resonant content – and if it’s good, it won’t matter if its synthetic, whether animated or photo-realistic. The illusion only breaks if it is genuinely poorly produced – and that’s the biggest worry about synthetic content at scale; most of it is unlikely to be good.”
That feels like the real tension. The question is not whether synthetic content can work. It can. The question is whether a marketing ecosystem obsessed with speed, efficiency and scale will use it to create more meaningful work, or simply more work.
From Content Creation to Content Production
The Operators Creative
The phrase “content creation” has always carried a little romance. It implies craft, thought, intention, voice, perspective, even when the actual output is a 400-word blog post about printer cartridges or a mildly desperate LinkedIn carousel about leadership.
“Content production” feels colder, but it’s probably more accurate now.
The generative AI era has industrialised the first draft. It has also industrialised variation. A single idea can become 50 headlines, 12 scripts, six visual territories, three tone-of-voice routes, a landing page, a newsletter, a product description, a podcast summary and a month of social content before lunch. Marketing teams are no longer limited by the same production constraints. They’re limited by judgement, brand discipline and whether anyone actually wants the content being made.
That shift is happening across a wider working culture where generative AI adoption is still uneven but growing quickly. A 2026 analysis of more than 36,000 workers across 35 European countries found average generative AI adoption at work was 12%, but ranged from under 3% to 25% depending on the country. Adoption was highest where workers had the skills, task autonomy and organisational conditions to use it meaningfully.
In other words, we’re not yet in a world where every workplace has fully integrated AI into production. But we are very much in a world where the content supply chain is changing.
That matters because production used to enforce selectivity. If making a video cost a meaningful amount of time and money, someone had to ask whether it was worth making. If translating a campaign into multiple languages required significant budget, there had to be a plan. If visual content needed photographers, illustrators, animators or designers, there was at least some friction between impulse and output.
AI removes much of that friction. That can be liberating, but it can also be dangerous. When content is cheap to make, businesses can forget to ask whether it deserves to exist.
This is where Manchipp’s warning about perfection becoming worthless starts to bite. AI is very good at producing surface polish. It can make things look expensive, resolved and professional. But if everyone has access to the same kind of polish, polish stops being a differentiator. The average rises. The middle gets shinier. The internet fills with competent-looking things that leave almost no emotional trace.
That’s why the move from creation to production is not just operational. It’s cultural. We’re seeing a shift from content as authored expression to content as generated output. From “what do we need to say?” to “what can we produce?” From editorial judgement to volume logic.
The risk for brands is obvious. They may become incredibly efficient at making things nobody remembers.
Why Synthetic Content Is Growing So Quickly
Eugene Borodulin
The growth of synthetic content is not mysterious. It’s happening because the incentives are almost irresistible.
Businesses want more assets, in more formats, for more channels, in more languages, at lower cost. Platforms reward constant presence. Search rewards freshness, authority and breadth, even as AI search reshapes how that content is discovered. Social channels reward speed, iteration and native format. Paid media requires endless variations for testing. Sales teams need enablement. Product teams need documentation. Employer brands need stories. Internal teams need comms. Audiences expect personalisation. Global markets expect localisation.
That is a lot of demand.
AI slots neatly into that pressure because it promises scale without proportional headcount. It can accelerate research, ideation, drafting, editing, versioning, translation, image generation, video prototyping and asset adaptation. It’s not hard to see why businesses are investing in it. For marketers under pressure to do more with less, AI generated content looks less like a toy and more like survival infrastructure.
The rise of AI search adds another layer. Brands are no longer only trying to rank in traditional search results; they’re also trying to be cited, summarised and represented accurately by AI systems. A 2026 study of Google Search, Gemini and AI Overviews found that AI Overviews were generated for 51.5% of representative real-user queries in its dataset and that source retrieval differed substantially between traditional search and generative systems. That means content now has to serve human readers, search engines and AI-mediated discovery environments.
This creates a strange feedback loop. AI creates more content, which makes discovery harder, which makes brands create even more content to remain visible. The web becomes a hall of mirrors, with synthetic material competing to be read by systems that may themselves produce synthetic summaries.
That is part of the reason the “dead internet” idea has entered the mainstream conversation. Stripped of its conspiratorial edges, the more useful version is simply this: more online spaces are being filled by automated content, bot activity, algorithmic feeds and generative media, making it harder to distinguish human expression from machine-produced output.
But it is important not to overstate the case. Synthetic content is not one thing. A badly generated spam article is synthetic content. So is an AI-translated customer help guide that makes support more accessible. So is a virtual production environment. So is an animated campaign using machine-assisted workflows. So is a personalised product recommendation. So is an AI-generated variant of an ad created under human direction.
That’s why Da Costa Gomez’s nuance matters:
“There are, however, some interesting areas where AI can allow us to deliver in meaningful ways at a grander scale than before. Personalisation, localisation, and multiple languages, for example, are all elements that can extend campaigns, and reach new audiences.”
That is the optimistic case for synthetic content. Not more sludge. More access. More relevance. More reach. More ability to adapt an idea across languages, audiences and contexts without rebuilding everything from scratch.
The danger is that too many brands will confuse the ability to make more content with the ability to create more value.
When Scale Becomes the Strategy
i2i Art Inc.
Scale used to be a distribution question. Make one strong piece of creative work, then scale it through media, PR, partnerships, search, social and culture. Now scale increasingly begins at production. Brands can generate more versions, more formats, more messages and more visual executions before anything has even reached an audience.
That changes the strategic conversation. Sometimes for the better. Often not.
At its best, scale allows a strong idea to become more useful. A campaign can be localised without losing its core. A product video can be adapted for different audience needs. A report can become searchable content, social clips, sales material and translated insight. A brand character can respond in more contexts. A complex service can be explained through personalised journeys.
At its worst, scale becomes the strategy itself. The business starts producing because it can. Content calendars fill up with machine-assisted filler. Search pages are targeted with generic articles. Social feeds become a wash of motivational phrasing, bland insight and suspiciously perfect imagery. The brand appears busy but not necessarily alive.
The evidence suggests audiences can sense this tension. One 2026 study of AI-generated content on a large Chinese video platform found that AI-generated creators achieved aggregate engagement comparable to human creators mainly through higher-volume production, despite users showing a marked preference for human-generated content. That is a fascinating and slightly grim finding. It suggests volume can compete with preference, at least in aggregate, which creates a powerful incentive to keep producing.
That’s how ecosystems get distorted. If high-volume synthetic content can perform acceptably, platforms and brands may push more of it into circulation. Not because it’s better, but because there’s more of it. The outcome is a scale-over-preference dynamic: audiences may prefer human work, but machine-assisted volume fills the feed anyway.
This is where Manchipp’s “digital landfill” line becomes more than colourful phrasing. It describes a structural problem. If content is cheap enough, some businesses will keep publishing until the environment itself becomes worse.
The same issue is already visible in advertising. Generative AI is being used to produce ads faster and more cheaply, while consumers have shown mixed reactions to obviously AI-generated signals, especially when visuals feel distracting or unnatural. The lesson is not that AI ads can’t work. The lesson is that scale and emotional resonance are very different things.
Da Costa Gomez makes that distinction beautifully. People will respond to good work, she argues, whether synthetic, animated or photorealistic. But “the illusion only breaks if it is genuinely poorly produced.” The real worry, as she says, is that “most of it is unlikely to be good.”
That should be pinned to every AI content strategy. The danger is not synthetic content. The danger is bad synthetic content at scale.
The Benefits and Risks of a Synthetic Content Ecosystem
Alfie Turnell
There is a version of the synthetic content ecosystem that could be genuinely useful.
It could make marketing more accessible for smaller businesses. It could help charities and public bodies communicate more effectively with limited resources. It could make content more inclusive through translation, captioning, audio versions and accessibility adaptations. It could help brands create more relevant journeys for different audiences. It could reduce wasted production effort. It could allow creative teams to prototype faster and spend more time on the ideas that deserve investment.
It could also help global campaigns feel less lazily global. Localisation has too often meant translation with a thin dusting of cultural awareness. AI-assisted workflows, when guided properly by local experts and creative teams, could make campaigns more responsive to language, context and audience need.
That is where Da Costa Gomez’s point about “personalisation, localisation, and multiple languages” is so important. Synthetic content has real potential when it extends the reach of a strong idea rather than replacing the need for one.
But the risks are just as real.
The first risk is quality collapse. If businesses use AI to produce more average content, the internet becomes more average. Not dramatically worse all at once, just flatter. Less surprising. Less specific. Less human. More likely to say the statistically expected thing in the statistically expected tone.
The second risk is trust. If audiences suspect content is generated without care, they may start discounting it before they even engage. Synthetic media also complicates questions of authenticity, disclosure and provenance, particularly when images, voices and likenesses are generated or manipulated. The UK’s regulator recently required Google to give publishers more control over whether their content is used in AI search features while preserving traditional search visibility, reflecting how quickly AI-mediated content use has become a live policy issue.
The third risk is discoverability. Search engines can penalise low-value, scaled content whether human or machine-made. Google’s long-standing position is that automation is not inherently against its policies, but using automation primarily to manipulate search rankings is considered spam. Its emphasis is on helpful, reliable, people-first content rather than the tool used to create it.
That distinction matters. The question is not simply whether search engines can distinguish between human and AI generated content. Detection is imperfect, and AI-assisted workflows blur the boundary anyway. The more important question is whether the content is useful, original, trustworthy and clearly made for people. Search engines may not always know exactly how something was made, but they are increasingly motivated to demote low-value scaled content because it damages the search experience.
The fourth risk is misinformation and error. AI-generated summaries and outputs can be inconsistent, especially in high-stakes areas. A 2025 audit of AI Overviews and Featured Snippets for baby care and pregnancy queries found inconsistencies between AI Overviews and Featured Snippets on the same search page in 33% of cases, while medical safeguards were present in only 11% of AI Overview responses in the study. That should make any serious brand cautious about using synthetic content in sensitive categories without robust human review.
The fifth risk is sameness. AI systems are trained on what already exists, which means they are often good at reproducing category norms. That can be useful for drafting, but disastrous for distinction. If every brand uses similar tools, similar prompts and similar benchmarks, synthetic content could accelerate blandification.
Manchipp’s answer to that is not more perfection, but more pulse:
“The new luxury is imperfection. The new premium is a pulse. If it looks hand made, and even better if there's proof it is, you'll make more of an impact.”
That is not nostalgia. It is positioning. In a world where flawless synthetic polish becomes cheap, evidence of human presence may become a premium signal.
What Happens When Everyone Can Publish Everything?
Torpedo
The internet has always been defined by abundance, but AI generated content changes the nature of that abundance. It does not merely make publishing easier. It makes production effectively elastic.
One person can now operate like a content department. A small business can behave like a media company. A bad actor can flood search results. A brand can generate endless variants. A creator can build synthetic characters. A publisher can automate summaries. A platform can populate feeds with generated material. A spammer can produce sites at scale. A marketer can fill every gap in the calendar with something that looks finished.
So, what happens when everyone can publish everything?
First, attention gets more expensive. Not necessarily in media cost alone, but in creative effort. It becomes harder to earn genuine attention because the feed is fuller, the baseline is more polished and audiences become better at ignoring anything that smells generic.
Second, provenance becomes more important. People will want to know where things came from, who made them and whether there is a real human or institution behind them. This may not matter for every piece of content, but it matters for brands built on trust, craft, expertise and emotional connection.
Third, distribution becomes more competitive. If content supply explodes, discovery systems become gatekeepers. Search engines, social algorithms, AI assistants and recommendation systems decide what gets surfaced. Brands that treat content as isolated assets will struggle. Brands that build coherent ecosystems of expertise, authority and audience trust will have a better chance.
Fourth, taste becomes more valuable. Anyone can generate options. Fewer people can decide what is worth publishing. This is where creative leadership becomes essential. The future content strategist may look less like a calendar manager and more like an editor, curator and cultural filter.
Fifth, human weirdness starts to matter more. The little flaw, the strange phrase, the risky reference, the awkward truth, the lived-in detail, the hand-made texture, the decision that feels too specific to have been averaged out by a model. These are the things synthetic content often smooths away. They may also be the things audiences start to seek out.
This is why Manchipp’s line about imperfection lands so well. He is not saying work should be sloppy. He is saying that in an over-polished synthetic environment, humanity becomes a differentiator.
Da Costa Gomez’s view is equally important because she does not reduce the debate to human good, synthetic bad. The audience response depends on emotional resonance. A synthetic piece can work if it has craft, care and feeling behind it. An entirely human-made piece can fail if it is dull, lazy or emotionally empty.
That is the standard brands need to apply. Not “was this made by AI?” but “does this deserve attention?”
The Future of Content in an AI-Driven World
Mehmet Turan
The future of content will not be human-only. That particular ship has not only sailed; it’s already been rendered, localised, clipped into vertical video and turned into 27 newsletter subject lines.
Nor will the future be fully synthetic, at least not if brands want to be remembered, trusted and liked. The more likely future is hybrid. Human-led ideas, AI-assisted production. Human judgement, synthetic variation. Human taste, machine speed. Human story, scalable adaptation.
But the hierarchy matters. If AI leads and humans merely approve, brands risk drifting towards the middle. If humans lead and AI extends, the tools become more interesting.
The content that will retain value in a world flooded with synthetic material will have several qualities.
It will have a point of view. Generic advice will become less valuable because generic advice is exactly what AI can produce endlessly. Brands need sharper opinions, clearer positions and more recognisable intellectual territory.
It will have evidence. Claims need sources, data, experience and proof. As AI-generated summaries and synthetic content multiply, unsupported assertions will feel cheaper.
It will have authorship. Whether that means a named expert, a visible creative process, a real community, a founder voice or proof of human craft, people will look for signs that the content comes from somewhere real.
It will have emotional intelligence. Da Costa Gomez is right that people respond to emotionally resonant content. The medium matters less than the feeling, provided the illusion holds and the work is genuinely good.
It will have restraint. The ability to publish everything should not become the decision to publish everything. Editorial discipline will matter more, not less.
It will have distribution thinking built in. In an AI-driven content ecosystem, the best content will not just be made. It will be structured, repurposed and surfaced intelligently across search, social, AI discovery, sales conversations, communities and owned channels.
And it will have a pulse.
That may be the simplest way to put it. The internet does not need more flawless emptiness. It needs more work that feels like it came from a mind, a culture, a team, a place, a tension, a risk, a joke, a memory, a belief.
AI generated content is not going away. Synthetic content will become more sophisticated, more common and harder to detect. Some of it will be brilliant. Some of it will be useful. A lot of it will be landfill.
The creative opportunity is not to reject the tools outright. It is to use them without becoming them.
That feels like the real challenge for creatives. Agencies, brands and creatives are not being asked to choose between humanity and technology. They are being asked to understand what each is for. AI can produce scale. Humans still need to provide taste. AI can generate variants. Humans still need to decide what matters. AI can simulate polish. Humans still need to create meaning.
In the end, synthetic content will force the industry to become more honest about value. If the work only existed because it was cheap to produce, AI will produce more of it. If the work existed because it had insight, feeling, tension, craft and point of view, it will matter even more.
The future will not reward the brands that publish the most.
It will reward the ones that know what is worth making in the first place.