Struck Off for AI Hallucinations: What the First SDT Ruling Means for Your Business

A registered foreign lawyer has been struck off by the Solicitors Disciplinary Tribunal (SDT) following a disciplinary hearing in which he relied on AI-generated quotes and citations that turned out to be false. The case, decided on 4 September 2026, marks the first time the SDT has directly addressed the issue of AI hallucinations — and it signals a watershed moment for professional services firms everywhere.

This was not a case about using AI. It was a case about using AI carelessly, failing to verify its output, and then presenting fabricated material as fact in a formal legal proceeding. The professional consequences were terminal. For accountants, solicitors, HR consultancies, and marketing agencies that have integrated AI tools into their workflows, the message is unambiguous: ignorance of what your AI produces is no longer a defence.

What Happened and Why It Matters

The details are instructive. The lawyer in question used AI-generated content — citations and quotations — in his own defence before a professional regulatory body. Those citations did not exist. This is a textbook example of an AI hallucination: a confident, plausible-sounding output that is factually wrong.

What elevates this case beyond a cautionary tale about one individual is the institutional response. The SDT treated the submission of false AI-generated material as a disciplinary matter serious enough to warrant striking off. That is the highest sanction available. The tribunal did not accept unfamiliarity with AI as mitigation. The professional was responsible for what they submitted, regardless of how it was produced.

For any firm where staff are using AI to draft documents, prepare client-facing materials, generate research, or support regulatory submissions, this ruling reframes the risk landscape entirely.

The Verification Obligation Is Now Real

Professional services businesses have been slow to formalise AI governance, often treating it as an IT concern rather than a compliance one. That position is increasingly difficult to sustain.

If a solicitor can be struck off for submitting unverified AI output in a tribunal setting, the logical extension is clear: any professional who presents AI-generated content as accurate — to clients, regulators, courts, or counterparties — carries personal and organisational liability for that content. The tool does not share responsibility. The professional does.

This means firms need documented verification protocols. Who checks AI-generated content before it leaves the business? What standard of review is required? How are errors recorded and corrected? These are not aspirational governance questions. They are the practical infrastructure that separates defensible practice from negligent practice.

The ICO Enforcement Register: 216 Actions and Counting

On the same date as the SDT ruling, the Information Commissioner's Office published an updated enforcement register covering actions taken between January 2022 and August 2026 — a total of 216 enforcement actions against organisations.

This figure matters for AI compliance because data protection and AI governance are inseparable. AI systems are trained on data, operate on data, and generate outputs that often involve personal data. Organisations that have deployed AI tools without conducting proper data protection impact assessments, without appropriate data minimisation controls, or without adequate transparency to data subjects are exposed on multiple fronts simultaneously.

The ICO has consistently signalled that AI-related data protection failures are within scope of its enforcement powers. With 216 actions now indexed — and the pace of enforcement unlikely to slow — businesses should treat the ICO register not as a historical document but as a live indicator of regulatory intent.

EU AI Act: Mandatory Compliance Now Applies to UK Firms Selling into Europe

On 6 September 2026, the European Commission's AI Office and national regulatory bodies across the EU published fresh guidance clarifying the compliance pathway for the EU AI Act. Its prohibition tier — covering unacceptable-risk AI applications — and its high-risk tier are now mandatory across all EU member states.

For UK-based businesses, the critical point is jurisdictional reach. If your firm sells AI-enabled products or services into the EU market, the EU AI Act applies to you, regardless of where you are headquartered. This is not a European problem for European businesses. It is a global compliance obligation for any business with EU clients or EU-facing operations.

The high-risk tier covers AI systems used in areas including employment decisions, access to essential services, and legal processes — all of which are directly relevant to HR consultancies, legal practices, accountancy firms, and professional services businesses broadly. Firms in these sectors that are using AI to assist with recruitment screening, performance assessments, credit evaluations, or legal research need to understand whether their systems fall within regulated categories, and if so, what conformity obligations apply.

UK firms that have been operating on the assumption that Brexit insulated them from EU AI regulation should revisit that assumption with urgency.

The UK IPO's Strategic Direction: Technology Is Central

Also on 4 September, the UK Intellectual Property Office published its Corporate Plan for 2026-27, which sets out a modernisation roadmap including significant technological upgrades. While this may appear tangential to AI compliance, it is relevant context for professional services firms advising clients on IP matters involving AI-generated content — an area of active legal development across multiple jurisdictions.

Questions of IP ownership in AI-generated outputs remain unresolved in many legal systems. As the IPO moves to modernise its systems and services, firms that advise on IP should be building the capability to guide clients through an evolving regulatory picture that is moving quickly.

What This Means for Professional Services Firms Globally

Taken together, these four developments — the SDT ruling, the ICO enforcement register, the EU AI Act mandatory tiers, and the IPO's modernisation direction — point to a consistent regulatory trajectory. AI use in professional services is moving from an informal, discretionary practice to a formally regulated one. The question is no longer whether your firm uses AI. The question is whether your governance keeps pace with the obligations that use creates.

For firms in the UK, US, Canada, the EU, the Middle East, and Asia-Pacific, the practical implications are the same: you need documented AI policies, verification workflows, staff training, and a clear understanding of which regulations apply to your operations and your markets.

A lawyer was struck off because AI hallucinated and no one caught it. That is the benchmark event. Everything that follows in regulatory enforcement will reference this moment.

Work With Ops Intel

Ops Intel helps professional services businesses build AI compliance frameworks that are proportionate, practical, and aligned with the regulatory environments in which they operate — whether that is UK, EU, US, or multi-jurisdictional.

If your firm is using AI tools and does not yet have documented governance to match, now is the time to act. Visit www.opsintel.io to learn how we can support your compliance programme.