The course · 16 contact hours
Eight sessions, two conversations, one cumulative arc.
The eight core sessions follow a cumulative arc in four phases of two. They begin by decoding confident claims and then opening the black box, so that the machine's failures can be predicted rather than merely listed (Sessions 1–2); move to computational thinking as the way to direct an AI system precisely (Sessions 3–4); turn to verification — tracing claims to origin, treating provenance as a legal question, and drilling the authority check (Sessions 5–6); and close by applying all of it to legal work and to the professional duties that make verification mandatory (Sessions 7–8).
The arc
Four moves, in order
Each phase builds on the last: decode a claim and then the machine itself, so its failures can be predicted rather than listed; learn the thinking skills that direct it precisely; drill verification until it is a habit; and only then apply all of it to legal work and the duties that govern it.
01 · Sessions 1–2
Decode the claim and the machine
Evaluate any confident claim — a person's, a statistic's, a chatbot's — then open the black box, so every failure mode that follows is one you can predict rather than memorise.
02 · Sessions 3–4
Direct the machine
The four pillars of computational thinking, taught as prompt-craft: decompose, recognise patterns, abstract, and give the model a procedure.
03 · Sessions 5–6
Verify everything
Trace a claim to its origin, treat provenance as the legal question it is, and drill the four-step authority check until it is automatic.
04 · Sessions 7–8
Apply it to law
Long judgments, doctrine and comparison — then the professional duties that make verification mandatory rather than advisable.
| Block | Shape | Hours |
|---|---|---|
| Core sessions | 8 × 90 min (60 taught + 30 hands-on) | 12 |
| Conversations with Experts | 2 × 90 min | 3 |
| Evaluation | 1 × 60 min | 1 |
| Total | 16 |
The core sessions · 60 taught + 30 hands-on
The eight sessions
Each session takes one idea, runs it as an experiment on the class, shows the same effect in a model, proves it live on a real chatbot, and lands on a named skill — then hands the room thirty minutes to do it themselves.
Conversations with Experts
Four closing hours: two practitioners, then the evaluation
Two 90-minute conversations, each led by a different practitioner and tied to material the students have already been taught — so the conversation tests the frameworks rather than introducing a parallel syllabus. A 60-minute evaluation closes the course.
The closing four hours
Two conversations with practitioners, then the evaluation.
Two 90-minute Conversations with Experts sit alongside the core sessions, each led by a different practitioner and tied to material the students have already been taught — so the conversation tests the frameworks rather than introducing a parallel syllabus.
Each runs as roughly 20 minutes of framed opening from the practitioner, 40 minutes of structured conversation against the course's own frameworks, 20 minutes of student questions prepared in advance, and a 10-minute named takeaway.
Speakers are confirmed with the host institution for each cohort; the two themes above are the recommended defaults and can be substituted.
Conversation 01 · 90 min
AI in practice
A litigator, in-house counsel or legal-technology lead on what has actually changed in the work: what they use, what they refuse to use, what went wrong, and what they now check before anything leaves the office.
Builds on Sessions 6–7 · speaker TBC
Conversation 02 · 90 min
AI, regulation and the bench
A judge, regulator or policy scholar on governance, evidence and where the rules are heading — including what a court does with an artefact nobody can authenticate.
Builds on Sessions 5 and 8 · speaker TBC
Evaluation · 60 min
A dedicated 60-minute evaluation session closes the course, taken in class.
Two parts. An applied audit: a fresh extract seeded with the three classes of authority failure — one fabricated, one that says something materially different, one no longer good law — to be verified against supplied sources. And short-answer questions on the course's frameworks: the Ladder of Misinference, the failure modes and their mechanism, the four pillars, the four-step authority check, and the professional duties. The paper rewards demonstrated verification and judgment over recall.
2 × 90 min + 60 min = 4 hours
Note on format
Every session is built around a four-step mirror
Every session runs 90 minutes in two halves. The first 60 minutes are built around a four-step “mirror” pattern: an experiment run live on the class; an explanation of how the same effect shows up in an AI model; a live demonstration on a real chatbot; and a concrete, named takeaway skill. The final 30 minutes are a guided hands-on segment on the students' own laptops, with a timed run of show, a fixed deliverable and a verification checklist.
- 01ON THE CLASS
An experiment on the class
Everyone commits an answer on their phone; the room sees itself fooled.
- 02IN THE MODEL
The same effect in the model
Why the identical shortcut shows up in a language model.
- 03LIVE CHATBOT
A live demonstration
Proved on a real chatbot, on the projector, then and there.
- 04THE SKILL
A named takeaway skill
One concrete discipline for using AI well — and what to do next time.
Why all eight
A course, not a collection of lectures.
The sessions are built to stack. The first two hand you the decoder — how a confident claim is judged, framed, and inferred, on the human side and the machine side. The middle sessions teach you to direct a model and to understand what it is. Session six makes verification a habit. By the time the course reaches the law, you can decompose a judgment, interrogate a summary, and run an authority check the way you would supervise a junior's brief. Apart, each session is a useful ninety minutes; together, they are a complete method for working with AI in the room.
Eight sessions, and one habit to leave with.
Each session ends in thirty minutes of hands-on work whose deliverable feeds the assessed portfolio and the audit. The labs page collects every brief in one place.