Landmark Ruling · OLG Hamm
Who Blames the Bot? The OLG Hamm Ruling and the Reality of AI Liability in Professional Services
In the rush to deploy generative AI, a comforting myth has taken root among business leaders: “As long as we train our models on verified internal data, we are legally insulated from its hallucinations.”
A landmark ruling from the Higher Regional Court of Hamm (Oberlandesgericht Hamm, May 12, 2026, Az. 4 UKl 3/25) has shattered this illusion.
The court ruled that a business is fully liable under unfair competition laws for the false or misleading statements generated by its AI chatbot—even if the company can prove it fed the system entirely correct data. For the professional services sector, from boutique law firms and medical clinics to international management consultancies, this ruling marks a tectonic shift. It marks the end of the “AI did it” defense and demands a fundamental re-evaluation of how AI is integrated into high-stakes client workflows.
A business is fully liable for false statements generated by its AI chatbot — even when the model was trained exclusively on accurate internal data. The “AI did it” defence is dead.
The Case
When “Dr. Bot” invents a Resume
The legal dispute involved Aesthetify GmbH, a medical cosmetic clinic group fronted by two high-profile, social-media-famous doctors known as “Dr. Rick” and “Dr. Nick.” The company integrated an interactive AI chatbot on their website to streamline patient inquiries and appointment bookings.
When prospective patients asked the chatbot about the doctors’ credentials, the AI hallucinated. It repeatedly claimed they were officially recognized “Specialists in Plastic and Aesthetic Surgery” and “Specialists in Aesthetic Medicine”—titles requiring rigorous, official medical chamber certifications that the doctors did not hold.
The Consumer Association of North Rhine-Westphalia (Verbraucherzentrale NRW) sued the clinic for misleading commercial practices under the German Act Against Unfair Competition (UWG).
The Ruling
No “Third-Party” Shield for Code
The clinic’s defense was predictable: they argued that they had not actively published the falsehoods, nor did they intend to mislead. The false claims were an unprompted, erratic anomaly of the generative AI model—a technical hallucination beyond their direct operational control.
The OLG Hamm categorically rejected this defense. The court’s rationale establishes a critical legal framework for the era of corporate AI:
The AI is not a “Third Party”
An AI system deployed commercially by a business cannot be legally categorized as a detached “third party.” Consequently, companies cannot rely on the restricted liability privileges usually granted to platforms hosting external, user-generated content.
Strict Zurechnung (Attribution)
Because the AI is actively utilized as an extension of the business to interface with consumers and drive commercial decisions (like booking appointments), its outputs are fully attributed to the corporation.
Data Quality is Not a Legal Shield
The court specifically emphasized that proving a model was exclusively fed accurate corporate data does not absolve management of liability. If you launch the tool to the public, you own its outputs — flaws, hallucinations, and all.
If you launch the tool to the public, you own its outputs — flaws, hallucinations, and all.
Given the systemic implications for corporate AI usage across industries, the OLG Hamm has allowed a revision to the Federal Court of Justice (Bundesgerichtshof). However, the baseline precedent has been set.
Sector Implications
Why Professional Services are Uniquely Vulnerable
While this case originated in the medical retail space, its core logic applies perfectly to any professional service vertical where trust, certifications, and compliance dictate market survival.
In public affairs, legal advice, tax consultancy, and financial planning, expertise is the product. Unlike a retail company where an AI chatbot might misstate a product’s shipping dimensions, a professional services firm risks misstating regulatory changes, compliance deadlines, judicial precedents, or liability limits.
If a public affairs consultancy uses a client-facing chatbot that misinterprets an upcoming EU ESG directive or falsely details an MP’s voting record, the agency can no longer blame the LLM provider or claim they had an internal compliance check on the training data. Under the OLG Hamm framework, that hallucination is an immediate, actionable breach of trade practices.
The Right Architecture
Moving Beyond Keywords: The Architecture of Accountable AI
At Policy-Insider.AI, our core methodology has always resisted the hands-off, “black box” deployment of generative AI. This ruling validates exactly why automated, unmonitored text generation is a strategic liability for sophisticated teams.
Many traditional monitoring and advisory setups treat AI as a standalone platform where users are left to query an unguided algorithm. When the goal is moving from reactive tracking to proactive strategy, that architecture introduces immense reputational and legal risk.
True professional-grade AI intelligence requires a distinct architectural philosophy:
AI for Scale, Humans for Judgment
Raw data ingestion, cross-border scraping, translation, and initial clustering are perfectly suited for automated scaling. But synthesis and strategic output must remain tethered to strict, reproducible rules and expert human oversight.
Replacing Static Documents with Controlled Validation
In professional environments, static position papers and regulatory assessments shouldn’t simply be queried via open-ended chatbots. They need to be continuously tested against real-world regulatory shifts via a highly structured data pipeline that guarantees trace-back references to the original institutional sources.
For Leadership
The Bottom Line
The OLG Hamm ruling serves as an urgent wake-up call for compliance officers and executives. If your organization is deploying client-facing or public-facing generative AI tools, your risk matrix must change.
You cannot contractually or technologically outsource your brand’s regulatory liability to a machine. If your AI hallucinates a regulatory fact, a professional credential, or a legal boundary, your firm will stand trial alone.
To thrive in this environment, professional services must move past the novelty of generative chat and invest in robust, issue-driven intelligence architectures where AI handles the noise, but humans retain control of the signal.
The OLG Hamm (Oberlandesgericht Hamm / Higher Regional Court of Hamm) is the highest regional court of appeal for civil and commercial law in its district. For international readers, here is why its rulings carry immense regulatory weight:
Germany’s Largest Court: Out of the 24 Higher Regional Courts in Germany, Hamm is the largest by case volume, serving an economic region of nearly 9 million people.
The Professional Services Specialist: By law, the OLG Hamm has exclusive statewide (not nationwide) jurisdiction over legal disputes involving the Freie Berufe (regulated professional services, including lawyers, accountants, auditors, and tax consultants).
Why it Matters: In the German judicial hierarchy, these courts sit directly below the Federal Court of Justice. Because its rulings dictate compliance in Europe’s largest economic market, an AI liability ruling from the OLG Hamm sets the structural benchmark for corporate risk management across the entire EU.
Build a Compliant, High-Precision Policy Monitoring Architecture
Don’t rely on generic chatbots that expose your team to algorithmic risk. Policy-Insider.AI builds tailored pilots for complex regulatory landscapes — with the trace-back references and human-in-the-loop architecture that the OLG Hamm ruling now makes mandatory.
Book a 20-minute scoping call →No credit card required · Set up in minutes