EU AI Act & High-Risk Systems
Classification, provider versus deployer obligations, and what changes when an institution adapts a general-purpose model for a high-risk purpose.
AI Regulation · Ethics · Governance
AI now decides who gets credit, coverage, and capital. The harder question is who stays accountable.
I’m Adv. Dvir Meirovich, an AI regulatory attorney. I co-chair the AI Committee at the Israel Bar Association’s Central District, and I’ve trained more than 2,000 attorneys and judges to use AI responsibly. Today I advise institutions on the harder part: how to use AI for decisions that carry real weight, and how to stay accountable for them.
Co-Chair, AI Committee
Israel Bar Association, Central District
0+ attorneys trained
and District Court judges
Three ventures founded
Big Wisdom · The Esquire Method · VERUM
AI regulation, ethics & governance
EU AI Act · GDPR · financial-sector exposure
Profile
A practicing lawyer who works with AI, not just around it.
Dvir Meirovich is an AI regulatory attorney and Co-Chair of the Artificial Intelligence Committee at the Israel Bar Association’s Central District. He has spent close to a decade on the legal side of new technology. More and more, that work has narrowed to one problem: what happens when an institution lets AI make a decision it has to answer for.
Through Big Wisdom, the training company he founded, he has taught more than 2,000 Israeli attorneys, along with judges on the District Court bench, to use AI in their actual work. That background shapes how he advises. He has put these tools in the hands of thousands of professionals, so his guidance comes from practice, not theory.
His work covers AI-specific regulation like the EU AI Act, the automated-decision rules that run from GDPR Article 22 onward, and financial-sector supervision. The same question sits underneath all of it. When an institution lets a general-purpose model make a regulated decision, what obligations does it take on, and how does it prove the decision was made properly?
He doesn’t just advise on these questions. He has built around them, founding The Esquire Method for US attorneys and VERUM for the institutions that have to stand behind their AI decisions.
A model can produce a recommendation. Turning that recommendation into a decision an institution can defend is a legal act, and right now almost no one is documenting it as one. Dvir Meirovich
Areas of Focus
Expertise
The rules that apply once AI starts making decisions that carry real consequences.
AI in Credit & Insurance
The regulatory exposure created when automated systems decide who receives credit, or life and health coverage, and who answers for the outcome.
Automated Decisions & GDPR Art. 22
Rights to explanation, what meaningful human oversight actually means, and the limits of fully automated decisions.
Cross-Jurisdictional Regulation
Mapping obligations across the EU, the UK (FCA, Bank of England SS1/23), the US (Colorado SB 24-205, CFPB, NAIC), Canada (Quebec Law 25) and APAC (MAS, HKMA).
AI Governance & Auditability
Oversight frameworks, documentation standards, and evidentiary records built to hold up under regulators, and eventually under courts.
Responsible AI Adoption
Putting AI to work inside legal and compliance functions without crossing ethical or professional lines, shaped by years of training thousands of practitioners.
Global Reach
Counsel that travels across jurisdictions.
One AI system can answer to a dozen regulators at once. I map the obligations across the regimes that decide whether a decision stands.
Europe
EU AI Act · GDPR Art. 22
United Kingdom
FCA · Bank of England SS1/23
United States
CFPB · NAIC · Colorado SB 24-205
Canada
Quebec Law 25
Singapore
MAS
Hong Kong
HKMA
EU AI Act
GDPR Article 22
Colorado SB 24-205
Quebec Law 25
UK FCA
Bank of England SS1/23
MAS Singapore
HKMA Hong Kong
US CFPB
NAIC Model Bulletin
Writing & Research
Insights
Analysis on the regulation and responsible use of AI, written for the people who have to act on it, not just argue about it.
EU AI Act · Featured
The Provider Trap: how a bank can become an AI “provider” without ever deciding to
Read the analysisThe obligation doesn’t arrive with a contract or an announcement. It arrives the moment the model is put to a purpose it wasn’t built for.
Ventures
He doesn’t just advise. He builds.
Big Wisdom
AI training for the legal profession
The legal-tech training company he founded. Through it he has taught more than 2,000 Israeli attorneys, plus judges on the District Court bench, to use AI in everyday practice.
The Esquire Method
Ethical AI workflows for US attorneys
Attorney-led AI workflow products for US lawyers, built so they can put AI to work without putting their license at risk.
VERUM
Independent AI decision infrastructure
Independent infrastructure that produces a verifiable, tamper-evident record of how an AI-assisted decision was made, so institutions can prove it, regulators can examine it, and courts can rely on it.
Visit
Patent pending
Work Together
Advisory
I work with financial institutions, law firms, and the legal, risk, and compliance teams inside them, on the regulatory and ethical questions that AI creates. The goal is always the same: clear, defensible decisions while the rules are still being written.
Start a conversation- i.
Regulatory strategy & AI Act readiness
Where you stand today, what the obligations actually are, and the path to meeting them.
- ii.
Governance, oversight & documentation
Frameworks for human oversight and record-keeping that satisfy supervisors and survive review.
- iii.
AI training & enablement
Equipping legal and compliance teams to use AI confidently and ethically, drawn from years of training thousands of practitioners.
- iv.
Expert opinions & regulatory analysis
Written analysis on specific systems, products, and exposures.
- v.
Board & executive briefings
Translating fast-moving regulation into decisions leadership can make with confidence.
Contact
Let’s talk.
For advisory work, training, speaking, or press, the fastest way to reach me is by email.