The Unseen War for the Soul of Journalism
In the quiet corners of the internet, a strange and unsettling transformation has been taking place. It’s not the kind of thing you notice at first, but once you see it, you can’t unsee it. I’m talking about the faces. The faces of people who don’t exist, smiling at you from op-ed pages, their bios confidently claiming decades of experience in fields they’ve never studied, their LinkedIn profiles adorned with headshots that are nothing more than pixels arranged by a machine. These are the “fake experts,” the ghostwriters of the algorithm, and they have infiltrated the news ecosystem with a level of success that should terrify anyone who cares about the truth. It started as a trickle, a few AI-generated columns here and there, but it has grown into a flood. Newspapers around the world, desperate for content and short on budget, have been duped into publishing the work of phantom journalists. One such apparition was a fake freelancer named Margaux Blanchard, a name that appeared on pitch emails and eventually in real publications, pitching inventive news features that were entirely fabricated by artificial intelligence. She wasn’t real. Neither was the Thai police force’s “undercover” photo, which showed hardened detectives dressed as showgirls—a purely AI-manipulated image that fooled newsrooms across the globe. This is the reality we live in now, and it is a war. A war for the very definition of what it means to be a journalist, a war between human curiosity and machine-generated bilge. And for a long time, the good guys were losing.
The front lines of this conflict have been most vividly drawn by Press Gazette, the UK-based journalism trade publication that has been documenting this crisis with a series of investigations it aptly named “Reality Wars.” Their findings have been nothing short of alarming. They’ve uncovered entire networks of fake bylines, each with a history, a headshot, and a fabricated backstory, all generated in seconds by large language models. They’ve watched as reputable, established titles have been forced to issue humiliating corrections and retractions after publishing stories that contained paragraphs lifted from nowhere and bylines that belonged to no one. The scale of the problem is staggering. It’s not just about the occasional rogue blog post; it’s about the systematic erosion of trust. Every time a fake expert is given a platform, every time an AI-manipulated image is passed off as real photojournalism, the public’s faith in the media takes another hit. And in an era where “fake news” is a casual slur thrown at any story the powerful don’t like, the media cannot afford to hand its critics a smoking gun. The damage isn’t just reputational; it’s existential. If readers believe that everything they see might be the work of a bot, they will eventually stop trusting anything at all. The very fabric of informed society begins to unravel.
This is the stark reality that prompted a rather fascinating survey conducted by Press Gazette earlier this year. They asked more than 200 publishers, editors, and journalists a simple set of questions about the use of AI in their newsrooms. The results were a clear signal of an industry in distress. A staggering 95% of respondents agreed that AI-generated content should be clearly flagged to readers. There’s an appetite for transparency, a desire to be honest with the audience. But here’s the kicker: 64% of those same people admitted that their publication has absolutely no reliable way to check for unauthorized AI use. They know there’s a problem, and they want to fix it, but they’re flying blind. They are, in essence, driving a car at high speed without a windshield, hoping they don’t crash into a deepfake they didn’t see coming. The disconnect is glaring. The industry recognizes the threat, wants to address it, but lacks the technical means to do so. This is a vacuum that needs to be filled, and it’s a vacuum that a group of people with a deep understanding of both journalism and technology have been working to fill.
Enter Made by Humans. It’s a name that sounds more like a badge of honor than a software product, and that’s exactly the point. Press Gazette, led by its editor-in-chief Dominic Ponsford, has partnered with Flip-Pay, a Dublin-based enterprise technology company, to build a platform that gives newsrooms the one thing they’ve been missing: a way to fight back. This isn’t just another piece of software; it’s a shield designed to protect the integrity of the written word. The platform works by integrating industry-leading AI-detection technologies—specifically Pangram for text and Identifai for images, video, and audio—into a single, unified dashboard. Think of it as a digital bouncer for your publication, checking the ID of every piece of content before it gets through the door. It can analyze text at both the article and sentence level, identifying whether it was likely written by a human or generated by an AI. It can scan images and video for the subtle, often invisible artifacts that indicate manipulation. It can even listen to audio and catch cloned voices that sound disturbingly like the real thing.
But what makes Made by Humans so revolutionary is not just its detection capabilities; it’s how it integrates into the daily workflow of a journalist. It’s not a separate, clunky piece of software that requires a PhD in computer science to operate. Instead, it plugs directly into content management systems like WordPress. This means an editor can be writing their headline, looking at a story that has just been filed, and a warning will flash on their screen if the text shows signs of AI generation. It’s a frictionless, automated safety net. It allows the editor to do what they do best—exercise human judgment—while the machine does the heavy lifting of forensic analysis. It’s a tool designed to support human editorial judgment, not replace it. And it’s a tool that has already been battle-tested, with Press Gazette becoming the first publication to trial it directly inside its own CMS, setting the precedent for others to follow.
Dominic Ponsford has been characteristically blunt about the stakes. He describes this as a “Reality War” because the fundamental nature of our reality—what we see, hear, and read—is being contested. “We’ve seen hundreds of stories taken down by reputable titles after elements have been identified as misleading or AI-generated,” he says. “This tool will help to stop such articles before publication.” He points out that Press Gazette, like many reputable publishers, does use AI for certain tasks like transcriptions, research, and data analysis. But there is a hard line in the sand: human creativity, originality, and curiosity must underpin the journalism. “This technology helps verify that all our content is ‘human-made’,” he adds. It’s a subtle but crucial distinction. They’re not Luddites trying to ban AI outright. They are realists, understanding that AI is a tool that can be used for good or for ill, and they are determined to ensure that their side uses it responsibly.
Paul McCarthy-Brain, the CEO of Flip-Pay, echoes this sentiment. His company works closely with publishers worldwide and has seen firsthand how AI is “undermining their commercial authority.” He sees Made by Humans as a crucial solution, not just for ethical reasons but for survival. “It offers editors a clear evidence-based assessment with an audit trail,” he explains, noting that this audit trail can be particularly useful for supporting compliance with regulations like the EU’s Digital Services Act. It’s one thing to have a hunch that something is wrong; it’s another to have the forensic evidence to prove it. This isn’t just about catching the bad guys; it’s about protecting the publishers themselves from the catastrophic reputational damage that comes from publishing misleading content. A single fake story can destroy years of trust built up over decades, and in the digital age, trust is the most valuable currency a news organization has.
Of course, the platform is only as good as the technology that powers it. For text, Made by Humans uses Pangram, a state-of-the-art detection model developed by the company of the same name. Pangram’s latest iteration, Pangram 4, is incredibly accurate, claiming to detect content from major AI labs like OpenAI, Anthropic, Google, Meta, and DeepSeek, while generating only one false positive for every 10,000 pieces of text it analyzes. Max Spero, the CEO of Pangram, emphasizes the increasing sophistication of his model. “Pangram 4 also distinguishes between AI-polished human writing and full AI generation,” he explains. This is a vital feature, as it allows journalists who might be using AI for basic tasks like transcription or light editing to pass the test. It isn’t about punishing anyone for touching AI; it’s about identifying when the human voice has been completely replaced by a machine. On the visual and audio front, Made by Humans is powered by Identifai, a deepfake detection specialist whose research has been published in peer-reviewed IEEE and ACM journals. Its lead image model detects AI-generated images with an astounding 99.8% accuracy, and its audio model, designed to catch cloned and synthetic voices, achieves a 91.8% success rate.
The data coming out of Identifai is a sobering reminder of the scale of the challenge. They have logged around 10,000 deepfake incidents reported in English-language media or on social platforms since January 2020. But it’s the trajectory that’s truly terrifying. The frequency of these incidents is rising exponentially, from a mere four reported incidents in January 2020 to more than 3,000 in March 2026 alone. That’s not a linear increase; that’s a hockey-stick curve that goes straight up. Kate Burns, Identifai’s Vice President for EMEA, puts it perfectly: “Almost anyone can now create a deepfake, and the most dangerous ones look exactly like the real thing.” She highlights the fact that while many of these deepfakes are caught by hard-working fact-checkers and picture desks, editorial teams are on the front line. They are the last line of defense between the public and a flood of convincing lies. “Made by Humans supported by Identifai gives media and broadcasters the vital tools they need to ask the simple question ‘is this real?’ and greater peace of mind to publish.”
In the end, Made by Humans is more than just a piece of software. It’s a statement. It’s a declaration that the pursuit of truth is not a battle that can be fought with a keyboard and a prayer alone. It requires the disciplined application of technology, the courage to admit there’s a problem, and the wisdom to build tools that augment human capability rather than replace it. The platform offers publishers the ability to earn a “trust badge” to display on their websites, a small visual cue that signals to readers and advertisers that this is a newsroom that takes responsibility for what it publishes. In a world where a well-crafted AI prompt can produce an entire newsroom of fake content in the time it takes to make a cup of coffee, having a verified, human-made seal of approval might just be the most valuable feature of all. As the line between reality and fabrication continues to blur, Made by Humans stands as a digital lighthouse, a beacon of hope reminding us that amidst the noise and the bots and the phantom bylines, there are still real people, doing the hard, unglamorous, and absolutely essential work of holding power to account. The machine may have learned to write, but it will never learn to care. That’s a distinction we must fight to preserve.