Image credit: The Velvet Sundown on X
An AI journalist recently fooled some of the world’s leading media outlets, including Wired, Business Insider and Index on Censorship.
‘Margaux Blanchard’ doesn’t exist, but her articles were published anyway.
A few weeks later, a writer calling himself Andrew Frelon revealed to Medium that he was, in fact, Blanchard or rather, he created her. She’d been pitching, writing and negotiating with editors across the world with minimal human involvement.
He claimed that his goal was to test and show how far AI could mimic human journalism and how much of the media ecosystem would notice or care.
Most didn’t notice for some time, but when they did, they definitely cared. It’s caused a stir about the ethics of AI use, and at least six publications have since taken down AI-generated articles under the name Margaux Blanchard.
Whether or not ‘Andrew Frelon’ is a real journalist is another rabbit role.
There are three articles associated with him on Medium, each claiming responsibility for one AI rouse or another. First, the Velvet Sundown (more on that later), and second Margaux Blanchard. His profile image looks suspiciously perfect, and the word ‘Frelon’ literally translates to ‘hornet’ in French. Perhaps this is another sting-laden, AI trip.
In one article, ‘Frelon’ claims that he is protecting his identity, which explains the false name and image. He describes his ‘real’ background is working for web platforms on safety and policy issues, and that he specialises in non-technical red teaming of web-based products and systems, including generative AI systems.
Whichever the case, the reality is unclear.
As AI becomes increasingly undetectable, and our minds suspicious, this unnerving state of being unknowingly fooled, en masse and on repeat, may just become increasingly common.
Are AI agents the future of journalism?
According to ‘Frelon’, Margaux Blanchard was just one node in a system called Calista-1, a network of AI agents trained to hunt for story angles, seed ideas on expert-quote platforms and even conduct voice calls and video interviews using beta avatars.
Whether or not Frelon is real, whoever is behind him wants to emphasise AI as the logical endpoint. The writer paints a bleak picture in an article published on Medium:
“Every worst nightmare that you can one day imagine happening in news is already happening and has been for quite some time. The incentives are too aligned for it to go in any other direction now but forward,”
“Editors want speed and volume. Platforms want engagement. Clients want influence at lower cost. Human journalists are expensive, slow, sometimes adversarial, or even worse: unionized. Autonomous agents are tireless, compliant, and cheap. Who do you think is going to win in the end? Soon human journalists will become irrelevant.”
AI band ‘The Velvet Sundown’ racks up millions of Spotify streams
Frelon also claims Calista-1 was behind the sudden boom in AI band The Velvet Sundown’s global press coverage, as well as their Twitter account (though not the band itself).
The AI-generated act comes complete with invented backstories, promo images and country-folk style albums, describing themselves as ‘not quite human, not quite machine’.
This summer, The Velvet Sundown gained more than a million Spotify streams in mere weeks. At the time of writing, they have over 300,000 monthly listeners, and their most popular song has over three million streams.
A spokesperson for Spotify said the company doesn’t prioritise AI-generated music, however, it has previously been criticised for populating some playlists with stock music created by fake artists.
Industry bodies from the Ivors Academy to the BPI have since called for mandatory labelling of AI-generated tracks, arguing listeners deserve to know what they’re hearing, and artists deserve to know if their work has been used to train these systems.
Creative industries push back against copyright infringement
Britain’s creative industries continue to rally together.
Earlier this year, national media printed the ‘Make it Fair’ campaign across their front pages in a coordinated initiative to protect the creative economy.
This month, some of Britain’s most iconic artists – Paul McCartney, Kate Bush, Elton John – sent an open letter to Keir Starmer demanding protection for creators ahead of a planned UK-US tech pact.
They argue Labour’s proposals to let AI firms train on copyright-protected work without permission amount to ‘mass theft’ of an artist’s life’s work. They point to international human rights covenants, the Berne Convention and the European Convention on Human Rights as grounds for legal protection.
Margaux Blanchard and the Velvet Sundown are emblematic of the ‘post-reality’ environment we now inhabit – one whereby AI agents, synthetic personas and AI-generated content intertwine with human work, often undetected. When they are exposed, they become stories in their own right and their reach grows.
The implications point to a future where trust, authorship and authenticity must be constantly interrogated.
Welcome to the post-reality era.
Navigating this new landscape demands strategy, creativity and transparency. At Brandnation, we create creative campaigns that cut through, and multiply impact across PR, influencer marketing, social media and performance marketing.
Interested to find out how we can help your brand thrive in the age of AI?
About the author
Natalie Clement | Digital
Marketing Executive
With international experience as a digital marketer, writer, and editor, Natalie has worked across sectors including lifestyle, technology, and tourism.
