It’s hard to overestimate the panic people are feeling around AI rules right now. This summer, an industrial designer I know wondered whether he had to tell a client he’d used AI to research a new product idea, and a construction company client who went all-in on AI across half of its processes mere months ago quietly began walking that initiative back, for fear it had crossed a line. These leaders weren’t confused about the ethics of using AI. At least, not exactly. They were unsure whether they had to disclose that they had used it, and, if so, why? Would failing to disclose forever mark them with a Scarlet AI?
Can you blame them? The ground beneath these questions is shifting at the speed of light. In response to whipsawing public sentiment about AI’s trustworthiness, major brands like Aerie, Equinox and Almond Breeze ran anti-AI campaigns in early 2026. And several (like Apple TV’s hit series Pluribus, which added “This show was made by humans” to its end credits) are now performing “humanness” as a trust signal. LinkedIn has rolled out a “seems like AI slop” button, letting users flag and instantly remove posts they suspect were written by AI—a real about-face for a platform that spent two years quietly rewarding AI-assisted posting. Substack, meanwhile, added an AI-detection feature built by a company called Pangram, which can estimate how much of any given newsletter was written by a machine. And as of last month, the EU introduced its AI Act, which enforces real transparency rules: as of August 2, 2026, under Article 50, organizations operating within the EU can no longer bury whether their content is AI-generated in the fine print. They have to come clean in a much more visible way.
But investor Stanley Druckenmiller’s fully AI-generated op-ed in the Wall Street Journal recently blew the debate wide open. It wasn’t just that the Journal openly admitted the op-ed was AI-generated. What really rankled critics was that nobody felt the need to apologize. “Of course I used AI,” Druckenmiller said. “I write everything using AI now for the same reason I use a calculator when I do math problems.” Meanwhile, the Journal’s editorial-page editor publicly defended him, calling AI “a fact of modern life” and mocking critics as “media pharisees.”
“A client who went all-in on AI mere months ago is now quietly walking the initiative back for fear it had crossed a line.”
I feel the need to air my personal grievances about this brouhaha. Not long ago, AI was being hailed as the greatest thing since the invention of the internet. If your company didn’t adopt it ASAP, you’d have about as much currency as a Model T Ford. Fast forward five minutes in AI time, and we’re now being shamed for using it. Now I have to listen to the CEOs of major companies swearing that they’d never use AI to write—which, as the head of a small company with a lean team—prompts me to roll my eyes and mutter to myself, ‘Of course you never use AI. You probably have a huge team who use it for you.’ To add insult to injury, whereas Druckenmiller was celebrated for his candour and invited to “write” for the Journal again anytime, a lesser-known novelist named Mia Ballard was accused of using AI in her horror novel, and her publisher cancelled the book and left her reputation in shreds.
My sense of what’s going on in this moment tracks with one recent workplace survey I came across, which found that eighty-one percent of employees believe leadership plays by a different set of AI rules than everyone else does. The tools were sold as a great equalizer, and the backlash is landing hardest on the people who don’t have a large team (or one at all), a big name, or a legal department standing behind them.
The fact that this AI whiplash happened so suddenly—and that government regulation, platform policy and brand behaviour are all converging on the exact question at the same moment—is undoubtedly why so many smart people, who are dancing as fast and as smartly as they can to adapt to AI, feel so unmoored right now.
“I roll my eyes at corporate CEOs who swear they never use AI. Of course they don’t. They probably have a huge team who use it for them.”
So, what to do, besides freaking out and screaming into the void? Instead of waiting for a rulebook, a few organizations have decided to make their own rules. What’s interesting is that they all landed on different ones. In January, Hermès relaunched its website with illustrator Linda Merad’s hand-drawn seahorses, eels, and products woven together. Deliberately imperfect, they show a human hand at work. Last month it commissioned artists Sarah Martinon and animator Armand Béraud to create a fresh animated world of sea creatures and airplanes. The couture house didn’t issue a statement explaining any of this. It didn’t have to. Standing out singularly amidst all the hyper-perfect AI imagery, a hand-drawn illustration provides all the disclosure necessary.
Samsú, a boutique Irish hospitality brand specializing in luxury, off-grid “digital detox” cabins nestled in the Irish wilderness, hired an illustrator in West Cork to redraw every icon on its website. He spent three months on the task, using ink made from foraged seaweed burned into pure carbon and mixed with a liquid binder, rather than relying on an AI design tool like Claude Design. The founder called it “probably the least efficient decision I’ve ever made in business”, betting that the people who chose to stay at Samsú would notice the attention to detail, sustainability and care, and that it would matter to them.
Every, a media company built around covering AI, discloses its AI use the way most outlets credit a photographer: routinely, and without ceremony. They say plainly in their published editorial guidelines that every piece is touched by AI in some way, and when an image is AI-generated, they credit the AI model, the same way they’d credit a photographer for one that isn’t. A named human author always signs off on the final piece and stands behind it. Every writes for AI-agent readers as well as human ones, and some of its formats are built to be machine-readable. For an outlet whose audience is partly other AI systems, disclosure isn’t a confession—it’s just accurate labelling.
And then there’s Clay, a tech company, which recently published an internal AI writing policy, not for its customers, but for its own team. Clay has four rules: you have to stand behind every sentence you hand someone; writing is how you think, so don’t skip it; you should spend more time writing a document than you’re asking someone else to spend reading it; and longer isn’t better, since AI’s default is to pad.
Clay’s approach functions less as a disclosure policy than a quality bar. And there’s research behind that second rule: a not yet peer-reviewed paper at the University of Maryland recently found that AI-written essays are far less likely to advance an original argument than ones written entirely by humans—meaning that the writing you skip may be doing some of your actual thinking for you without you even noticing, which makes sense to me, because, like most writers, I write to find out what I think.
“Instead of waiting for a rulebook on AI disclosure, some organizations have decided to make their own rules. They all landed on different ones.”
What all of these brands have in common is that each worked out, specifically, its promise to its audience—craftsmanship, cutting-edge candour, rigour—then asked itself whether its use of AI supported or broke that promise. That’s a completely different question than, Am I allowed to use this astonishing new tool? Or, Should I feel ashamed if I do?
This matters whether you’re a commercial studio courting clients, or an independent contractor whose work depends on someone trusting you. For a purpose-driven organization, the stakes can be higher: a single social post that used AI for something as ordinary as scheduling or research can become a story about your organization betraying its own mission, even when you did nothing wrong. That gap between the brand promise and the behaviour is where reputations get hurt.
All of which is to say that you don’t need a hard-and-fast rule, because there isn’t one. You just need to think about the promise you’re making to the people who trust you. Your audience is buying a version of you they believe is real. What they actually want and expect from you is an informed take from a trustworthy source, much in the same way you rely on your doctor for a medical diagnosis, not the verdict you anxiously cobble together from late-night symptom-Googling. Once you understand that promise, you need to write it down, in plain language, before a platform, regulator, or Pangram outs you for using AI.
“There’s no hard-and-fast rule. Just think about the promise you’re making to the people who trust you, write it down in plain language and communicate it in your own way.”
Here’s how we use AI at Parcel: we rely on an AI-assisted workflow. For Unpacked, we rely on AI for brainstorming and refining ideas, preliminary research, structure and organization, and proofreading, but—and this is key—I’m the person flying that plane. I direct AI and then I check and correct it. I review, verify and stand behind every word that goes out under my name, and because I’m a professional with years of experience and a Spidey sense for spotting AI gaffes, I don’t feel that I have anything to hide—or apologize for. As I’ve said before, but I think bears repeating: AI is a tool, a fantastically useful, time-saving, scarily effective tool. Why wouldn’t I avail myself of its many wonders?
If you’re wondering where to begin, I suggest drafting a few concrete rules, like Clay’s four. Make them plain, specific, and about how you use AI, not how you feel about it in the abstract. Where you use it. Where you don’t. How you use it. Who checks and signs off on it. When you’re done, ask yourself if each of your answers matches the promise you’re making to the people who trust you. Then revisit that promise in six months (or six days) because here’s another promise: by then, the ground will have shifted beneath you again.
One Last Word
If you’re suffering from AI fatigue, as I am, I feel your pain. I long to write about other topics. But AI is the one people in my world are talking and fretting about most right now, for obvious reasons. So, barring the need for another emergency AI missive, here’s my final promise to you: next month, this space will be an AI-free zone.
