A PRACTICAL AI PRIMER FOR LAWYERS: THE SECOND EDITION OF “ANDREW AND THE MARVELLOUS ANALYTICAL ENGINE”: MAJOR ADDITIONS TO A FAST MOVING AND IMPORTANT AREA

November 14, 2026 · by · in Applications, Artificial Intelligence, Avoiding negligence claims, Civil Procedure

The use (and often misuse) of artificial intelligence has featured regularly in this blog over the past three years.  Initially it was a litigant in person, however a large number of practising lawyers have fallen foul of hallucinations. Further it has caused problems in expert evidence, witness evidence and even in judgments.  The fact is, however, it is here to stay.  It must be understood and properly controlled.  This second edition of Andrew Hogan’s book helps us all along the say.

“There are serious implications for the administration of justice and public confidence in the justice system if artificial intelligence is misused.”

Ayinde; Al-Haroun [2025] EWHC 1383 (Admin), paragraph 9

 

THE SECOND EDITION

If you are reading about this book and it sounds familiar, it is possible because I reviewed the first edition,  in September 2025. That  was a compact introduction that combined Hogan’s personal history of computing with demonstrations of AI tools. The 2026 edition runs to nearly 100,000 words and is organised around the work lawyers do and how they should check and supervise it.

 

THE MAJOR CHANGES BETWEEN THE EDITIONS

Apart from the fact that I have a hardback copy of the book, and there are a number of AI drawn pictures of Andrew and his Labrador, there are important key elements.

 

Area First edition (2025) Second edition (2026)
History and structure Hogan’s own experience, from the ZX Spectrum onwards, provides a route into the history of computing and AI. Compresses some of that history and gives more space to how modern AI works, how lawyers can use it, and how they should control it.
AI tools Demonstrations of ChatGPT, Copilot, Google’s AI tools and Perplexity, including how to use a custom GPT to draft legal articles. Removes standalone product chapters. Covers legal AI platforms and gives a dedicated chapter to AI agents that can carry out a sequence of tasks, alongside AI-assisted coding.
Day-to-day legal work Covers legal workflows and gives practical examples of using AI. Gives separate chapters to legal research, drafting and review, litigation workflows and practice management. Examples include chronologies, skeleton arguments, document analysis, costs budgets, bills of costs and funding agreements.
Professional duties and evidence Covers ethics and regulatory concerns. Expands the treatment of fabricated authorities, verification and supervision. Gives dedicated chapters to confidentiality, privilege and data governance, and to AI and evidence. Updates the case law and covers guidance from the SRA, Law Society and Bar Council.
Practical resources Worked example on drafting legal articles. The appendices include a model AI policy for chambers, further reading and a list of online resources current at publication.

THE CLB BOOK REVIEW CRITERIA

Regular readers know that there are basic issues we need to look at.

THE BOOK

Andrew and the Marvellous Analytical Engine: A Practical AI Primer for Lawyers.

WHO IS IT AIMED AT?

It’s aimed at practising lawyers, particularly solicitors, legal excutives and barristers who want to understand AI and use it responsibly in their day-to-day work. It assumes no background in coding or computer science.

WHAT DOES IT COVER ?

The book explains how AI works, how lawyers can use it for research, drafting, litigation and practice management, and what they must do to verify its output, protect confidentiality and meet their professional duties.

IS IT VALUE FOR MONEY?

Yes. It cost £9.99 Kindle, £24.99 paperback and £29.99 hardback. This is far cheaper than a telling off by the court and subsequent disciplinary proceedings…

SHOULD I BUY IT?

If you are aiming to be in practice in the future then you have to know these things. You cannot afford not to buy it. (Deliberate double negative if any pedants are reading this…)

WHAT IS MISSING?

It would have been useful to have a checklist/set of guidelines on how (in the short term) to avoid problems with Artificial Intelligence. However that may not be fair criticism, it may well the job of something like this site (so see this post).

WHERE CAN I BUY IT?

Here. 

RELATED POSTS ON ARTIFICIAL INTELLIGENCE

Date Post Brief summary
4 Nov 2026 More hallucination points to watch: AI can simply make things up AI added words to legislation. The post illustrates why practitioners should check the exact statutory text before relying on an apparently plausible quotation.
13 Oct 2026 Preventing problems with the use of artificial intelligence in litigation: 10 key points Draws together ten practice points from the blog’s coverage of AI use in litigation.
9 Oct 2026 Artificial intelligence hallucinations: causing problems for the profession again Examines another decision illustrating the procedural difficulties caused by reliance on AI-generated material.
7 Oct 2026 The authenticity of documents: ten key points on the risks posed by AI Practical points on checking provenance and dealing with apparently authentic documents that AI may have generated or altered.
7 Oct 2026 False documents, late evidence and artificial intelligence Looks at an Upper Tribunal decision involving late evidence and the risk that AI can produce convincing false documents.
1 Oct 2026 AI and civil justice: preparing for the tsunami Summarises Lord Briggs’ speech on AI, access to justice, procedural rules and the role of the human judge.
17 Sep 2026 When ChatGPT drafted a 300-page skeleton argument A 132,000-word skeleton exceeded a 20-page limit. The litigant remained responsible for complying with the rules.
15 Sep 2026 Lawyers and AI: avoiding the pitfalls (5) — pleadings and proof A Scottish case illustrates how AI can help litigants formulate allegations that they lack evidence to establish.
10 Sep 2026 Lawyers and AI: avoiding the pitfalls (4) — experts and the use of AI Twenty practice points on the risks when experts use AI in preparing reports or giving evidence.
9 Sep 2026 Lawyers and AI: avoiding the pitfalls (3) — AI and witness evidence Twenty points on AI-assisted witness statements, including the risk that drafting obscures the witness’s own recollection and language.
8 Sep 2026 Lawyers and AI: avoiding the pitfalls (2) — how to spot false authorities Covers invented citations, inaccurate propositions attributed to genuine authorities, and practical checks to spot both.
7 Sep 2026 Back to Basics Monday: lawyers and artificial intelligence (1) Sets out three principles on responsibility for the work, checking primary sources and verifying that an authority actually supports the proposition advanced.
7 Sep 2026 Hallucinated case law and legal principles: the Solicitors Disciplinary Tribunal Examines the SDT’s response to AI misuse in submissions defending disciplinary proceedings.
26 Aug 2026 Expert Watch 68: “Create me an exceptional expert witness report” A US case examines an expert’s prompts seeking a report that put the client at zero fault, raising questions about independence and reliability.
31 Jul 2026 Artificial intelligence in court: a practice direction from the High Court in Ireland Highlights Ireland’s High Court practice direction on generative AI in court documents and its relevance to practitioners elsewhere.
22 Jul 2026 Witness Evidence Wednesday: when a witness gets coaching from AI Explores how a criminal court dealt with evidence that a complainant had received detailed AI coaching.
13 Jul 2026 Hallucinated case law and the Crown Prosecution Service Examines the filing of CPS documents that relied on authorities that did not exist.
9 Jul 2026 The TCC guidance on the use of artificial intelligence Examines the AI section of the Technology and Construction Court Guide and its implications for checking material put before the court.
1 Jul 2026 Artificial intelligence and litigation: an update from the Civil Justice Council Reports the CJC’s preliminary consultation findings on AI-assisted preparation of court documents.
16 Jun 2026 Bar Standards Board guidance on AI and other technologies Summarises the BSB guidance and gives seven checklists for barristers using AI and related technology.
16 Jun 2026 SRA guidance on effective supervision (2): AI guidance and checklists Considers how the SRA’s effective-supervision guidance applies to AI use and provides practical checklists.
2 Jun 2026 Notice to admit facts: no deemed admission and a possibly hallucinated rule Explains the effect of silence in response to a notice to admit facts and examines a fictitious procedural rule that may have been AI-generated.
25 May 2026 The use of AI: landing solicitors in serious trouble again Examines solicitors’ reliance on false authorities even though the AI tool had warned them to verify its output.
15 May 2026 AI and misleading authorities: if your name is on the document, you own it A further case of misleading authorities illustrates the responsibility of the lawyer whose name appears on the court document.
30 Apr 2026 AI in the preparation of witness statements: how does the judge know it is the witness’s own words? The judge questioned whether a statement prepared with ChatGPT reflected the witness’s own account.
31 Mar 2026 More use of AI, more hallucinated cases A further false-citation case illustrates the care required when using AI for legal research.
31 Mar 2026 Guidance for those using AI to carry out legal research The Irish Court of Appeal considers responsible AI-assisted research and the obligation to verify authorities and propositions before relying on them.
23 Mar 2026 Hallucinations keep appearing in the reports: two more examples Reviews two examples of counsel putting forward hallucinated authorities, including one where counsel acted in their own case.
20 Feb 2026 AI and hallucinated cases: lawyers who delegate remain responsible The Upper Tribunal addresses responsibility for accuracy and supervision when legal research is delegated.
19 Feb 2026 The use of AI for preparing court documents: CJC interim report and consultation Introduces the Civil Justice Council’s interim report and consultation on using AI to prepare documents for court.
11 Feb 2026 Witness Evidence Wednesday: AI in the drafting of witness statements The court considers the difficulty of distinguishing a witness’s evidence from language or content supplied by an algorithm.
30 Jan 2026 AI generating phantom references and false quotations Authorities cited by a respondent failed to support the propositions advanced, and two did not contain the words attributed to them.
14 Jan 2026 Inaccurate case summaries: the representative who would “neither confirm or deny” AI use The First-tier Tribunal identified inaccurate summaries in written submissions and questioned how they had been prepared.
15 Dec 2025 False cases generated by AI: lessons from the courts of Oregon Uses an Oregon decision to discuss why fabricated citations threaten the administration of justice and why the term “hallucination” can understate the seriousness.
12 Dec 2025 Review of the Year (2): hallucinated cases in the courts Looks back at the year’s false-authority cases, court responses and what those developments might mean for lawyers using AI.
20 Nov 2025 Another false citations case: “I relied on the AI Overview” from Google Examines false citations traced to Google’s AI Overview and the need to check the underlying authorities.
23 Sep 2025 AI used in the writing of a judicial decision Looks at AI use in drafting a judicial decision and extends the discussion from advocates’ use of AI to judicial use.
19 Sep 2025 Book review: Andrew and the Marvellous Analytical Engine, by Andrew Hogan Reviews the first edition of Hogan’s book on AI for lawyers and its treatment of the opportunities and risks for legal practice.
15 Sep 2025 Barrister referred to the BSB over AI-hallucinated cases Examines a regulatory referral and counsel’s responsibility to check citations even where they were unaware of how the cases had been generated.
14 Jul 2025 Hallucinated cases led to success at first instance, then reversal on appeal An appellate court overturned a decision reached after reliance on false cases. A further false case was then cited in support of costs.
19 Jun 2025 Another case on AI and hallucinated cases: possible mischief in the background Reviews a further false-authority case and the court’s concerns about how the material came to be put forward.
9 Jun 2025 Citing false cases to the court (4): cases where AI caused trouble at home and abroad Summarises the Divisional Court’s review of domestic and overseas cases involving AI-generated authorities.
6 Jun 2025 Citing false cases to the court (3): the problems, guidance and penalties Extracts guidance from the Divisional Court judgment and discusses the potential consequences of putting false material before a court.
6 Jun 2025 Citing false cases to the court (2): lawyers must verify citations supplied by clients Covers the Al-Haroun proceedings and the responsibility to check authorities supplied by a lay client.
6 Jun 2025 Citing false cases to the court (1): the case of Ayinde Looks at the Divisional Court’s treatment of the false authorities in Ayinde and the question of contempt.
19 May 2025 The citation of false authorities: legal representatives have a date in court Reports the listing of the hearings concerning false authorities, ahead of the Divisional Court judgment.
7 May 2025 False cases in written arguments: wasted costs against counsel and solicitors Covers the first-instance wasted costs decision in Ayinde, where written arguments relied on five authorities that did not exist.
15 Oct 2024 When an expert tried to use AI to support their calculations Examines an expert’s use of Microsoft Copilot to support calculations and the resulting problems with verification and reliability.
10 Apr 2024 Artificial intelligence in the courts: selected highlights from the judicial guidance Selects points from the AI guidance for judicial office holders, including confidentiality and the risk of inaccurate or fabricated output.
7 Dec 2023 Using AI to search for case law and make submissions: it makes cases up – it really does In Harber v HMRC, a litigant in person relied on nine decisions that did not exist. The First-tier Tribunal concluded that the case summaries had been generated by AI.