Chances are you use AI to some extent. Whether it’s inbuilt AI-driven automation systems or a last-minute event flyer, almost every business in print, packaging, life sciences, manufacturing and technology now does.
Far fewer have read the regulation that, from 2 August 2026, governs how you tell people about it. Below, we share our practical guide to the EU AI Act, and why one small clause matters more to your marketing than the rest of the Act put together.
If you run a print, packaging, life sciences or manufacturing business, there’s a fair chance you’ve mentally filed ‘EU AI regulation; away. It sounds like something aimed at the tech giants in Brussels’ sights, not at a converter, a contract manufacturer, or a specialist ingredients supplier.
That’s an understandable interpretation, but it’s also a very risky one.
As of 2 August 2026, the transparency rules of the EU AI Act are live. They don’t care how big you are or what sector you’re in. They care about what your AI does, and crucially for anyone who publishes content, about whether you tell people when a machine wrote it.
Where is the EU AI Act now?
The EU AI Act (Regulation (EU) 2024/1689) is the first comprehensive AI law anywhere in the world. It first entered into force in August 2024 and has been phasing in ever since, sorting AI use into four risk tiers and attaching heavier obligations to the riskier ones.
Let’s take a look at the timeline to date:
- February 2025: The outright bans (prohibited practices) and the AI literacy obligation took effect.
- August 2025: Rules for general-purpose AI models (the ChatGPTs, Claudes and Geminis of the world) kicked in.
- 2 August 2026: The transparency obligations now apply. This is the one that could impact many ordinary businesses.
- December 2027 and August 2028: The demanding “high-risk” obligations, which were pushed back by a reform package (the Digital Omnibus) agreed earlier this year.
The scariest-sounding parts of the Act, conformity assessments, technical documentation, and EU database registration, have been deferred. The priority right now is transparency.
The EU AI Act risk tiering system.
The four tiers run from banned to barely regulated, and knowing where you sit saves a lot of needless worry.
Prohibited risk
Covers things the EU has decided are simply too harmful, such as social scoring, manipulative systems, and emotion recognition in workplaces. Almost no legitimate B2B business goes near this.
High risk
Covers AI that materially affects people’s rights or safety. Examples here include CV-screening and recruitment tools, credit scoring, biometric ID, and critical infrastructure. Unless you build or use tools like these, it probably isn’t you. And even if it is, the deadline has moved to late 2027.
Transparency (or limited) risk
This is the one that matters to most SME businesses. It’s less about what your AI does and more about whether you’re honest about using it. Chatbots, AI-generated content, and deepfakes. This is where the vast majority of businesses using everyday AI tools will land.
Minimal risk
This covers almost everything else, including spam filters, recommendation engines, and the AI features baked into software you already use. Unregulated, bar a general duty to make sure your people understand the tools they’re using.
For a typical print, packaging or life sciences business, the most likely answer is that you live in the transparency and minimal tiers.
The EU AI Act rules, in plain English.
The transparency obligations sit in a part of the Act called Article 50, and they break down into a handful of common-sense duties:
- If you run a chatbot or virtual assistant on your website or social media platforms, it has to tell people they’re talking to AI, upfront.
- AI-generated images, audio or video whether they’re on the website, or your social media channels, that could pass for real must be labelled.
- AI-generated content more broadly needs to be marked in a machine-readable way so it’s detectable as synthetic.
- AI-generated text published to inform the public on matters of public interest has to be disclosed as AI-generated.
Two things make this bigger than it first appears. First, ‘matters of public interest’ is broad. It isn’t just news reporting, it reaches into health, financial, scientific, environmental and regulatory developments. It’s precisely the territory that serious B2B thought leadership occupies. A blog explaining what PPWR means for your customers, or how a substrate change affects recyclability, is close to this line.
Secondly, these rules apply to deployers, not just the AI companies. A deployer is anyone using an AI system in their business. You, in other words, not just OpenAI or Anthropic. If your team generates content, runs a branded chatbot, or publishes AI-assisted copy, the obligation is yours to manage.
Being in the UK doesn't get you off the hook.
If you’re reading this from Britain and feeling relieved, don’t. The Act follows the AI output, not the company’s postcode. If your content or systems reach an audience in the EU, be it EU customers, EU website visitors, or sales into EU markets, the standards still apply. If that logic feels familiar, it should. It’s the same extraterritorial reach GDPR has had for years.
Plus, given the UK’s track record of eventually mirroring EU digital regulation, getting your house in order now is less about this deadline than about not scrambling to retrofit everything when a domestic version inevitably arrives.
Picture a flexible packaging supplier using AI to sharpen a product shot for a LinkedIn graphic, smoothing a crease, perfecting the gloss, until the pack looks flawlessly, believably real. That’s AI-manipulated imagery presented as authentic, and it’s precisely what the marking and labelling rules exist for. The supplier isn’t thinking about the AI Act; it’s thinking about a nice-looking social media post.
Or a machine manufacturer that hands its new brochure copy to a generative tool to save a fortnight of writing. If that copy stays purely promotional, it may well sit outside Article 50. But the moment it starts explaining throughput gains, emissions performance or compliance credentials, it edges toward the public-interest territory the disclosure rules cover. And either way, unreviewed AI copy in a technical sector tends to read exactly like everyone else’s.
Neither business set out to create a compliance problem, but both could, without ever realising it. This is exactly where a marketing partner, who understands both the rules and the market, is irreplaceable. A team of experienced industry professionals, like Think B2B Marketing, knows where the line sits and makes sure your content lands on the right side of it.
How the human touch is rewarded.
There’s one particular clause that may change how you think about all of this.
The obligation to label AI-generated public-interest text falls away where the content has been through genuine human review and editorial control, with a real person or organisation – like Think B2B Marketing – taking editorial responsibility for it.
The regulation has effectively written a marketing principle into law. Content that a human expert has genuinely authored, shaped and stands behind sidesteps the disclosure requirement. Content churned out by a machine and pushed live unchecked does not.
The bar is deliberately high. Regulators have been clear that a token human ‘check’ doesn’t qualify. The draft guidance points toward a documented editorial process with named, accountable people. You can’t launder unreviewed AI output through a thirty-second skim and call it editorial oversight. It’s why the value of an experienced marketing team, who not only produce razor-sharp content through experienced minds, but can also bring an editorial eye to existing AI content, makes enormous sense.
We’ve been making the case for years that in technically demanding markets, real human expertise is what cuts through. That when everyone reaches for the same AI tools, the output converges into a homogenous, forgettable sameness. In sectors where credibility is the entire game, sameness is a major commercial risk.
Buyers can tell when it’s AI. Journalists can tell. Google’s EEAT signals are built to tell.
Now the law tells too. The businesses producing genuine, expert-led, human-owned content don’t just solidify that all-important trust, they also stay comfortably on the right side of a regulation that’s about to put a ‘synthetic’ label on everyone who didn’t bother.
It’s not only regulators tightening up. Search engines now reward genuine, people-first expertise and demote the thin, repetitive content that AI churn tends to produce. So overdoing it doesn’t just dent your credibility, it costs you visibility. A growing share of our work at Think B2B Marketing involves exactly this: using experienced human hands to rebuild websites, marketing collateral and content strategies that went overboard on AI.
What the EU AI Act means in practice.
You don’t need to panic, and you certainly don’t need to swear off AI. The sensible response is straightforward. Work out which tier you’re actually in. For most businesses, that’s transparency and minimal. Make sure any customer-facing chatbot discloses itself, and check that the content you publish has genuine editorial ownership behind it, not just a machine and a rubber stamp.
But the salient point is that the AI Act has turned something we’ve always argued into something you can now be fined for getting wrong. The human voice was already the one that stood out in a crowded market. From 2 August, it’s also the compliant one.
That’s the part no tool can do for you, and it’s exactly where we can step in to help. If you want to talk through what the AI Act means for your content and your credibility, get in touch with the Think B2B Marketing team at hello@thinkb2bmarketing.com.
Frequently Asked Questions
The EU AI Act first entered into force in August 2024 and applies in phases. Its transparency obligations apply from 2 August 2026, while the more demanding high-risk obligations have been deferred to December 2027 and August 2028.
It can. The Act follows the AI output rather than the company’s location, so a UK business whose content, systems or products reach an audience in the EU can fall within scope. It’s the same extraterritorial reach that GDPR already has.
Article 50 sets out the Act’s transparency obligations. It requires businesses to tell people when they’re interacting with AI (such as a chatbot), to label deepfakes and synthetic media, and to disclose AI-generated text published to inform the public on matters of public interest.
For chatbots, deepfakes and realistic AI-generated images, audio or video, yes. For AI-generated text on public-interest matters, you must disclose it — unless it has been through genuine human review and editorial control, with a named person or organisation taking editorial responsibility.
Purely promotional copy may sit outside the disclosure rules, but content that informs the public on regulatory, technical, health or environmental matters is more likely to be flagged. The cleanest way to stay on the right side of the line is genuine human authorship and editorial ownership.
Breaches of the transparency obligations can attract fines of up to €15 million or 3% of global annual turnover, whichever is higher. Using a prohibited AI practice carries the steepest penalty, up to €35 million or 7% of global turnover.
