Law firms, legal tech companies and AI labs now need lawyers who can judge whether an AI's legal work is right. This programme turns your legal training into that skill, in plain English.
Self-paced · 10 chapters and a capstone · Certificate on completion
Prompt: "Breach of a supply contract in Scotland. What is the time limit?"
"Claims must be brought within six years of the breach, as confirmed in Harrow Developments v Pemberton [2018]. Time can be extended where facts were deliberately concealed."
JUR · CriticalSix years is the England and Wales rule. Scotland uses a five-year prescriptive period.CIT-FAB · CriticalAuthority cannot be found in any primary source.OKConcealment point is directionally right; flag for the Scottish equivalent.AI evaluation is not engineering. Engineers build the system; evaluators decide whether its legal output is right, safe and usable. That decision needs a lawyer, not a programmer.
The rules don't say "hire an AI evaluator". They demand human oversight, documented testing and lawyers who stay accountable for every output. Evaluation is how firms meet them, and evaluators are the people who produce the evidence.
High-risk systems need risk management, technical documentation, logging and human oversight. Following the Digital Omnibus, these obligations apply from 2 December 2027.[5]
Breaching the Act's obligations for high-risk systems can bring fines of up to €15 million or 3% of worldwide annual turnover, whichever is higher.[6] The same regime covers high-risk AI in healthcare, which is why the Career Accelerator places you on a live healthcare AI project.
In Ayinde v Haringey (2025), the High Court found existing guidance insufficient and referred lawyers to their regulators over fake citations.[9]
Completing this programme does not make you or your firm compliant with any regulation, and it is not endorsed by any regulator. It teaches the evaluation skills that compliance work depends on.
A new professional role at the meeting point of law and technology, already advertised by AI labs and growing inside firms and legal tech companies.
Written for practising lawyers, with worked legal examples, interactive exercises, charts and a checkpoint quiz in every chapter. Every fact is referenced so you can check it.
This is not a simulation. You will evaluate a healthcare AI platform that is in beta testing now, funded by SynaptiQ Tech and Riders for Health. It is built to bring AI-supported diagnosis to people in Africa who would otherwise have no access to a doctor.
When an AI talks directly to patients, every answer carries two kinds of risk. You will learn to judge both.
OutputExposure map and scope memo
OutputTest set, rubric and error log
OutputFindings report and exposure register
Because you work on a live product, you can list the project on your CV as work experience in AI evaluation, backed by your findings report and badge.
Graduates can volunteer on the project periodically, on an ongoing basis, keeping their experience current while helping bring safer care to more people.
Both tracks include the complete course. The Career Accelerator adds the hands-on project, peer review and a verifiable badge.
For junior lawyers and individual professionals
For ambitious associates and job-seekers
Training a team? Group bookings for firms and legal departments are available on request: email us.
AI evaluation specialist, computing lecturer and Lead User Researcher
Phemi earned his MSc in Architecture: Computing and Design from the University of East London in 2005, specialising in artificial life algorithms and artificial intelligence, and has worked across digital, UX and teaching ever since. For eight years he was a lecturer in computing at University Campus Milton Keynes.
He has served as Head of UX Research in a UK Government department, with more than ten years of research experience across UK government. His specialism is AI and LLM evaluation and human-AI interaction research, backed by over twenty years of software development. He designed this programme to give lawyers the evaluation method without the technical barrier.
Yes. The course explains how AI models work in plain English, with diagrams and demonstrations. Every exercise uses legal judgement: reading, marking and writing. You will never be asked to write code.
Solicitors, barristers, paralegals, in-house counsel and legal academics who want to work in AI evaluation, legal tech quality or AI assurance, either alongside practice or as a career move.
No course can do that on its own, and this one is not endorsed by any regulator. It teaches the evaluation skills that compliance work depends on: testing AI systems against clear criteria, recording evidence and explaining risk.
Worked examples use the law of England and Wales, with comparisons to Scotland and the US to teach jurisdiction errors. The regulatory material covers the SRA, the courts of England and Wales and the EU AI Act.
Yes. Chapters 1 to 3 cover the role, how AI models work and where legal AI fails. Read them before deciding which track suits you.
A live healthcare AI platform in beta testing, funded by SynaptiQ Tech and Riders for Health, that brings AI-supported diagnosis to people in Africa who would otherwise have no access. You evaluate both patient safety and the legal exposure created by how the AI interacts with people, working under the project's confidentiality and data protection rules.
Healthcare AI is regulated as high-risk, and when it talks directly to patients, questions of consent, data protection, scope of advice and liability arise in every conversation. That makes it the toughest practice ground there is, and the skills transfer directly to legal tech and other regulated sectors.
Core Track learners receive a completion certificate. Career Accelerator learners also receive a verifiable badge, a portfolio of evidence, and work experience on a live project that they can list on their CV. Neither is a regulatory qualification.
Yes, if you choose. Career Accelerator graduates can volunteer on the project periodically, on an ongoing basis. It is optional, flexible, and a way to keep your evaluation experience current.
No coding. No technical background. Three chapters free, then choose the track that fits your career.