The sanctions have escalated since this was written. On 27 July 2026 the Raad van Discipline in 's-Hertogenbosch imposed a reprimand and EUR 1,250 in costs on a lawyer who put non-existent ECLI numbers and citations to unrelated cases into briefs in two matters. The council said openly that it had stepped back from a suspension because AI literacy "was not yet commonplace" in 2025. That mitigation has an expiry date attached to it, which is the argument this piece makes. Separately, the AI Act's transparency obligations under Article 50 became applicable on 2 August 2026, with a grace period to 2 December 2026 for marking content produced by systems already on the market.
The Dutch regulator just punished three lawyers for submitting AI-fabricated case law. Marketing runs the same risk every day without consequence. That gap is the grace period, and it is closing faster than most CMOs realise.
Earlier this year, the Netherlands' regulator for the legal profession issued formal warnings to three lawyers. Their offence: submitting court documents containing citations and case law that did not exist. The fabrications were generated by AI. The lawyers submitted the output without verification. The court noticed. Two of the three were also ordered onto an AI training course. Dutch business daily Het Financieele Dagblad covered one of the cases in detail.
If you are a marketing leader reading this, enjoy your grace period. It has a shelf life.
The lawyers were punished because law has a verification culture. Every citation has a source. Every case has a docket. The profession has spent centuries building infrastructure around a single question: is this claim true, and can I prove where it came from?
Marketing has not built that infrastructure. A fabricated statistic in a keynote deck is a talking point. A hallucinated customer insight in a campaign brief is a narrative. A confident-sounding claim in a product page is copy. The same flawed AI output that gets a lawyer sanctioned would, in marketing, get a round of applause for being "sharp".
This is the grace period. It is not a stable condition.
Why the grace period will end
Three forces are compressing it.
- Regulation is travelling outward from high-stakes professions. Law went first because courts are the natural pressure test. Healthcare, financial services and HR tech are next. Marketing sits downstream of all three. The moment an AI-generated ad triggers a consumer protection inquiry, or an AI-personalised healthcare campaign misstates efficacy, the verification burden arrives.
- Platforms are starting to audit. Ad systems are already enforcing against AI-generated misrepresentation. Social platforms are labelling AI content. The EU AI Act's transparency obligations land at the content layer, not just the model layer. The audit trail marketers have avoided for twenty years is becoming a product requirement.
- Buyers are learning to spot it. B2B buyers in particular are developing pattern recognition for low-substance AI content. The polished blog post that says nothing is starting to generate the same response as the cold email that opens with "I hope this finds you well": instant filter, immediate distrust.
What this looks like through Value Gravity™
The Value Gravity™ model argues that the visible AI layer, where the text, images and copilots live, is low-mass. Value does not stick there. It flows downward, towards the dense, governed, high-switching-cost base: CRM, CDP, identity, data quality, policy, audit trail.
The Dutch lawyer story is a Value Gravity™ story told in a different vocabulary. The AI output looked fine at the top layer. The legal system cared about what sat beneath it: source, provenance, verification. That lower layer is where accountability lives, and it is where value, in the long run, always accumulates.
Marketing's coming reckoning will follow the same logic. The winners will not be the teams producing the most AI content. The winners will be the teams with a governed base underneath: clean first-party data, verifiable customer insight, source-of-truth content repositories, audit trails that show what a model was trained on and what it actually produced.
That is not a top-layer problem. That is an architecture problem.
A note on why I went deeper
I recently completed Anthropic's AI Fluency certification. The course deliberately avoids the prompt-engineering theatre. It focuses on how models reason, where they break, and why they fail in predictable ways. The short version: the gap between using AI and understanding AI is where the expensive mistakes live.
The Dutch lawyers were on the wrong side of that gap. Most marketing teams I meet right now are there too. They just have not been tested yet.
What to do before the grace period ends
A short list for CMOs and MarTech leaders who would rather not be the cautionary tale at next year's conferences:
- Treat AI output as a draft, not a deliverable. Build a verification step into every AI-assisted workflow. It does not need to be heavy. It needs to exist.
- Invest in the layers beneath the AI. Data quality, identity resolution, governed content libraries, clean CRM records. This is where generative AI either becomes reliable or becomes a liability.
- Document what your AI is doing. Which tool, which model, which prompt, which data. Future you, or a future regulator, will want to know.
- Train for judgement, not for tools. The marketing profession spent a decade chasing certifications in platforms. The next decade's valuable skill is the ability to read AI output critically and recognise when it is confidently wrong.
The closing thought
Law is not a more serious profession than marketing. It is a more audited profession than marketing. That gap is closing. The two-year window to build verification discipline, data quality and governance into marketing operations is not a luxury. It is the last quiet period before the audits arrive.
The grace period is a gift. Do not spend it on another AI-generated blog post.
The Gravity Scan maps your marketing stack across 28 assessment areas, identifying where governance, data quality and verification discipline are strong enough to carry AI workflows, and where they are not.
Frequently asked questions
What happened in the Dutch lawyer AI case?
Regional supervisors for the legal profession issued formal warnings to three lawyers who submitted court documents containing AI-generated citations and case law that did not exist. Two of the three were also ordered onto an AI training course. Sanctions have escalated since: on 27 July 2026 a disciplinary council imposed a reprimand and costs in a separate matter.
Why does this matter for marketing teams?
Law has a verification culture with documented sources for every claim. Marketing does not. Fabricated statistics, hallucinated customer insights and confident-sounding AI copy currently carry no regulatory consequence, but regulation is moving outward from high-stakes professions, platforms are auditing, and buyers are learning to spot low-substance AI content.
What should CMOs do before the AI grace period ends?
Treat AI output as a draft rather than a deliverable and build a verification step into every AI-assisted workflow. Invest in the layers beneath the AI: data quality, identity resolution, governed content libraries and clean CRM records. Document which tool, model, prompt and data underpin each output. Train teams for judgement on AI output, not just tool proficiency.
When do the EU AI Act transparency rules apply?
Article 50 transparency obligations became applicable on 2 August 2026 and were not deferred by the Digital Omnibus on AI. A four-month grace period runs to 2 December 2026 for the machine-readable marking duty on generative systems already placed on the market before 2 August 2026. The high-risk obligations under Annex III were the part that moved, to 2 December 2027.