From this August, the way businesses use GenAI will be regulated under Section 50 of the EU AI Act. These rules will have little or no impact on how marketers use AI. That doesn’t mean, however, that it’s not time to shake things up.
EU AI Act Section 50: What marketers need to know
The Section 50 transparency rules obligate businesses to disclose to users when they’re interacting with AI agents. Brands must also clearly label content that’s been created by GenAI.
Given that 66% of marketing professionals worldwide use AI tools and that at 55% text-based content creation is the leading use case, you’d be forgiven for thinking that Section 50 will lead to a seismic shift in how marketers use GenAI for copy. Nothing could be further from the truth.
That’s because the rules come with two significant exceptions. AI disclosures are not required if 1) the text has undergone a process of human review or editorial control and 2) a human or company holds editorial responsibility for the final published text.
None of what I’m saying should be construed as legal advice, but it seems clear to me that as long as you don’t simply put out AI-generated copy verbatim you’re good from a compliance perspective. In effect, most marketers will be able to carry on as before (at least in so far as copy is concerned; AI-generated images and voice are another matter and beyond the scope of this blog).
Should marketers use AI-generated copy?
However, just because businesses can continue to generate content using AI (so long as there’s a human in the loop) it doesn’t automatically follow that they should. Here’s why:
- AI doesn’t save that much time. Ninety-three percent of marketers edit AI-generated content before publishing – as well you might when served up generic, characterless copy. Editing and fact checking is such a time-consuming job that in many cases it’s simpler and faster to just write the copy yourself. Doing so also ensures that your content is unique.
- AI can negatively affect AEO. Content that has had significant human input outperforms AI-generated content in AI summaries. Similar to the above point: if you’re going to have to spend time editing content just to make it original, why not just write it from scratch?
- Media outlets reject AI-generated content. Earned thought leadership is more important than ever, as it shows up well in AI search summaries. However, publications increasingly ask contributors to formally declare their content has not been written by AI. In these situations there’s no room for doubt: you must write all your own copy if you’re to secure the placement.
- Using GenAI makes you dumber. Evidence suggests that the more you use AI, the more it may negatively affect your critical thinking skills. The very act of writing helps you to think. If you stop using this capability you risk losing it. That’s bad for you, and for the brands you work for.
Finding a balance between human and machine
AI can be a useful tool for copywriting. I find it particularly useful for understanding what not to write. Because AI summarizes the consensus view (it essentially repeats what most people are saying on a given topic) it can be used smartly to identify narrative “white spaces” – unrepresented thought leadership areas that your brand can “own.” However, marketers should confine their use of GenAI as much as possible to research and use it rarely, if at all, for actual writing.
Unless you’ve been using AI to automatically generate entire streams of content, the EU AI Act will have little impact on you. However, that doesn’t mean you shouldn’t change the way you work. High-quality writers that craft truly original content will always have the edge over machines. In a digital world dominated by AI slop, brands that swim against the tide and use human writers will stand out.




