GDPR fines have crossed €5.88 billion cumulatively, and the enforcement priorities have shifted. Regulators are now chasing dark patterns, AI-driven processing, and healthcare data. If your outbound stack touches any of those, the legislation that governs you just got sharper. The EU AI Act's August 2026 deadline adds a second layer of obligation on top.
Here's the practitioner read: the penalty headline is noise. The enforcement focus is the signal.
The €5.88B number is a distraction. The target list is the story.
Every quarter someone posts the cumulative GDPR fine total like it's a scoreboard. It isn't useful. What matters is where regulators are pointing the telescope right now.
Three priorities have hardened over the last twelve months:
- Dark patterns. Consent flows that nudge, pre-tick, or bury the opt-out. This is now a first-tier enforcement category, not a footnote.
- AI processing. Any automated decision-making that touches a prospect's data without a lawful basis and a documented logic trail.
- Healthcare and special-category data. If your ICP includes clinics, healthtech, or anything adjacent, your margin for error is thinner than you think.
None of those are abstract. If you're running outbound to EU prospects with an AI tool that scores inboxes, drafts emails, and enriches contact data, you are inside the blast radius of at least two of those categories.
What the AI Act actually adds on top of GDPR
The EU AI Act's August 2026 compliance deadline is not a GDPR sequel. It's a parallel obligation. GDPR governs personal data. The AI Act governs the system that processes it.
For most outbound teams, the practical question is whether your tooling qualifies as a high-risk AI system. The honest answer for a standard cold email workflow is usually no. But the moment you add automated scoring that materially affects a person's access to a service, or you're selling into regulated sectors, the classification can flip.
What that means in practice:
You need a documented record of what your AI tool does with prospect data, what model or logic it uses to score or rank, and who is accountable when it gets something wrong. That's not a legal team deliverable. It's a founder deliverable. If you can't answer those three questions in a paragraph, you have a gap.
We've written before about how AI scrutiny is the real threat to cold email, not the penalty. That piece still holds. The AI Act just makes the scrutiny formal.
Dark patterns are the quiet killer for outbound landing pages
Most outbound teams obsess over the email and ignore the landing page. That's backwards under the current enforcement climate.
If your demo-booking page has a pre-ticked consent box, a confusing unsubscribe flow, or a cookie banner that makes "accept all" a big green button and "reject" a grey link, you are running a dark pattern. Regulators have started treating those as standalone violations, independent of whether the underlying data processing was lawful.
A customer running outbound to UK accountancy firms found this out the hard way. Their email copy was clean. Their landing page had a pre-ticked marketing consent box inherited from a template. That single checkbox was the exposure, not the 4,000 emails they'd sent that quarter.
Audit your landing pages before you audit your sequences. It's faster and it's where the fines actually land.
The legislation stack is now three layers deep
If you're selling into the EU, the UK, and the US, you're not managing one law. You're managing a stack:
Layer one: GDPR and UK GDPR. Lawful basis, consent, data subject rights. This is the foundation and it hasn't changed.
Layer two: the EU AI Act. System-level obligations on top of the data-level ones. New as of August 2026.
Layer three: national privacy laws. Twenty US states with twenty different rules, plus Australia's shifting exemption threshold. We covered the single compliance rule that works across all twenty US states if you need the shortcut.
The mistake founders make is treating these as separate projects. They're not. One documented data flow, one consent record, one AI accountability note covers all three layers if you build it once and keep it current.
What we changed in our own stack
We're not a law firm and this isn't legal advice. But here's what we actually did after the enforcement priorities shifted.
First, we mapped every place prospect data enters our system: form fills, enrichment tools, inbox scoring, CRM sync. One page, one diagram. It took an afternoon and it surfaced two tools we'd forgotten were still connected.
Second, we rewrote our landing page consent flow. No pre-ticked boxes. Reject is as easy as accept. It cost us a small percentage of opt-ins and it removed a category of risk we couldn't quantify.
Third, we wrote a one-paragraph AI accountability note for each tool that scores or drafts. What it does, what data it sees, who owns the output. That note now lives next to the tool in our stack doc.
Fourth, we kept the human-in-the-loop rule absolute. Nothing sends without a person pressing the button. That's not just a product principle. Under the AI Act, it's the cleanest defence you have that a human, not a system, made the decision.
What we'd do next
If you're running outbound into the EU or UK, spend one afternoon on the data flow diagram and one on the landing page audit. Those two artefacts cover most of what the current enforcement wave is actually chasing. The AI Act deadline is close enough that the documentation should exist now, not in July.
If you want a tool that keeps the human in the loop by design and never sends without your finger on the button, give MiraReach a try.
— Mira