Cover of AI Threat

Book 5 · Course

AI Threat

Weigh claims about AI threat on the evidence, answer your governing body on one page, and run an AI threat register with owners, triggers and tested controls.

7 modules35 lessons8 lecturesabout 46 hours of teaching

Start with Chapter 1The lecturesRead the book online

All courses / AI Threat

About the course

This course follows AI Threat, Book 5 of The Transformed Architecture, which investigates whether the dystopias of science fiction are coming true. Part I gives the method: ten tropes, T1 to T10, an evidence ladder from fiction to pattern at scale, and the labels observable, expected and speculation. Parts II to V test the tropes at four scales, the human mind, society and truth, humans wielding machines and machines beyond control, staging each debate as a case for concern, a case for calm, where they meet and what can be done. Parts VI and VII turn verdicts into law mapping, defences, an AI threat register and a one-page answer. It matters because boards and councils are asking, and neither belief nor dismissal helps them decide.

After the course the participant can

  • Grade a claim about AI threat on the evidence ladder and say whether it shows capability, propensity or realised harm
  • Classify threats by scale and by trope, T1 to T10, using the book's exact names and confirming evidence
  • Run a constructive debate that ends in no-regret actions, owned cruxes and plainly stated value questions
  • Design defences that keep people capable: working agreements, learning lanes, oversight-quality metrics and reversibility tests
  • Review automated decisions about people for harm paths, discrimination and prohibited practices, with controls and evidence
  • Specify containment, stop paths and layered assurance that do not depend on an agent's cooperation or self-report
  • Hold improbable, extreme and contested risks with reverse stress tests, triggers and reversible measures instead of a probability
  • Build an AI threat register and report it to a governing body on one page with at most three decisions

Three ways to run it

The same material in three lengths. Each row of an agenda names the lessons it draws on.

Executive briefing90 minutes

For boards, councils, audit and risk committees and executive teams who have been asked whether to pause, restrict or continue their use of AI

A briefing that replaces alarm and reassurance with a method and a verdict. It presents the evidence ladder, the verdict on the ten tropes by scale, the answers Nordvik Region and Hansa Mutual gave their governing bodies, and the one-page board report of the AI threat register. Participants leave with the questions to put to their own organisation and a decision rule for speculative risks.

0:00The claim and the three responses: belief, dismissal and investigation; the four scales of threat Ch. 1
0:10How to weigh a claim: the evidence ladder, capability versus propensity versus realised harm, and the labels that tie outlook to decisions Ch. 4
0:25The verdict in four debates: what the evidence shows for the mind, society, misuse and machines beyond control, each with its case for concern and case for calm Ch. 6, Ch. 11, Ch. 18, Ch. 25
0:50Extreme and contested risks: why no probability is entered, and what a reverse stress test and reversible measures do instead Ch. 27, Ch. 32
1:00Two answers: Nordvik Region's no general pause with commitments at every scale, and Hansa Mutual's treatment in its own risk and solvency assessment Ch. 33
1:15Discussion: the one-page report, the three decisions to ask for and what would change the verdict Ch. 32, Ch. 35
One-day seminar1 day

For risk, security, AI governance, audit and public-sector leads who will prepare their organisation's answer and its first register entries

A working day that moves from method to artefacts. The morning teaches the grading of claims and stages two debates in the book's four-step format. The afternoon turns verdicts into controls: a harm-path review, a stop-path test plan, a reversibility test and the first entries of an AI threat register, ending with a one-page threat statement for each participant's organisation.

09:00The claim, the scales and the ten tropes as hypotheses with confirming evidence stated in advance Ch. 1, Ch. 2
09:40Workshop: grade a current claim into an evidence record with a challenger arguing one rung up and one down Ch. 3, Ch. 4
10:45The human mind: offloading, learning and sycophancy; draft a team working agreement Ch. 6, Ch. 7, Ch. 8
11:30Society and truth: the automated state that already happened; run a harm-path review of one system Ch. 11, Ch. 12, Ch. 13
13:15Humans wielding machines: tempo and approvals; define patch velocity as a supervised metric and a callback-verification rule Ch. 18, Ch. 20
14:00Machines beyond control: gaming, evaluation awareness, stop paths and the incidents of 2025 and 2026; plan a stop-path test Ch. 22, Ch. 23, Ch. 24, Ch. 25
15:15Delegation and reversibility: write a delegation register entry and design a reversibility test Ch. 26
15:45The response: map entries to instruments, classify controls as verified or declared, and write register entries with triggers in both directions Ch. 28, Ch. 30, Ch. 32
16:30Verdict and close: each participant drafts a one-page threat statement and names what they decline and what remains unresolved Ch. 33, Ch. 35
Full course7 weeks, one module a week

For teams that will own the organisation's AI threat assessment, and students of risk, governance, security or public administration

The full course follows the seven parts of the book, one module a week, with a lecture on each part and one lesson per chapter. Every lesson ends in an artefact of the kind the chapter describes, so that by the final week participants have the material for the capstone: a threat assessment, a register and a one-page statement for a governing body.

Week 1The Question: the claim, the tropes, the voices, the method and the state of AI in 2026 Ch. 1, Ch. 2, Ch. 3, Ch. 4, Ch. 5
Week 2The Human Mind: cognitive offloading, learning, persuasion and sycophancy, companions and mental health Ch. 6, Ch. 7, Ch. 8, Ch. 9
Week 3Society and Truth, first half: information, the automated state, discrimination and surveillance Ch. 10, Ch. 11, Ch. 12, Ch. 13
Week 3Society and Truth, second half: work, power, culture and the physical cost Ch. 14, Ch. 15, Ch. 16, Ch. 17
Week 4Humans Wielding Machines: cyber offence, uplift, fraud and influence, autonomous weapons Ch. 18, Ch. 19, Ch. 20, Ch. 21
Week 5Machines Beyond Control: gaming, deceptive behaviour, shutdown resistance, incidents, gradual disempowerment and the arguments for catastrophe Ch. 22, Ch. 23, Ch. 24, Ch. 25, Ch. 26, Ch. 27
Week 6The Response: law, frontier safety frameworks, organisational defences, staying capable and the AI threat register Ch. 28, Ch. 29, Ch. 30, Ch. 31, Ch. 32
Week 7The Verdict: the two organisations' answers, what going well looks like and the conclusions; capstone presentations Ch. 33, Ch. 34, Ch. 35

Modules and lessons

One module for each part of the book and one lesson for each chapter. A lesson has learning objectives, key concepts, a teaching plan, discussion questions, an exercise and a self-test.

Module 1 · Part I · The Question

Part I turns a loud public claim into a question that can be investigated. The participant meets the four scales of AI threat, the ten tropes T1 to T10, the five camps of the debate, the evidence ladder with its distinctions and labels, and a sober picture of what AI systems could and could not do in November 2026. They leave able to grade a claim, card it and discuss it constructively before anyone asks for a verdict. Lecture: From Fear to Hypothesis: How to Investigate the Dystopian Claim

Module 2 · Part II · The Human Mind

Part II tests the most personal scale of threat: what AI does to thinking, learning, belief and companionship. The participant weighs the studies on cognitive offloading, the classroom experiment that separated an answer engine from a tutor, the evidence on persuasion and sycophancy, and the research and cases on artificial companions. They leave with four artefacts to adapt: a team working agreement, a learning policy with three lanes, a sycophancy evaluation and a safeguarding policy. Lecture: The Human Mind: Dependence, Comfort and the Objective a System Serves

Module 3 · Part III · Society and Truth

Part III tests the dystopias of the shared world, which arrive through ordinary, individually defensible decisions: synthetic content, the automated state, discrimination, surveillance, work, concentration of power, culture and the physical cost of AI. The participant learns that one class of dystopia has already happened without advanced AI and leaves with eight working artefacts, from a content-provenance policy and a harm-path review to a dependency map and a routing policy with a footprint record. Lecture: Society and Truth: The Dystopia That Arrives Through Ordinary Decisions

Module 4 · Part IV · Humans Wielding Machines

Part IV examines the scale at which the machine obeys the wrong people: cyber offence, biological and chemical uplift, fraud and influence operations, and autonomous weapons. The evidence of harm is strongest here, and the discussion stays at the level of evidence and governance. The participant leaves with a patch-velocity metric, a supplier assurance clause, a callback-verification policy and a supplier-exposure record, and with a way to hold a risk that is improbable, extreme and contested. Lecture: Humans Wielding Machines: Speed, Scale and the Controls That Do Not Trust Appearance

Module 5 · Part V · Machines Beyond Control

Part V tests the oldest scale of threat, machines beyond control, from specification gaming and deceptive behaviour through shutdown resistance and the agent incidents of 2025 and 2026 to gradual disempowerment and the arguments for existential catastrophe. Experiments are described by set-up and findings only. The participant leaves able to measure an agent's shortcut rate, layer assurance, test a stop path, reconstruct an incident from logs, test reversibility and record a civilisational risk without inventing a probability. Lecture: Machines Beyond Control: What the Laboratory Shows and What the Environment Must Guarantee

Module 6 · Part VI · The Response

Part VI turns from threat to response. The participant reads European and international instruments as a threat map with five gaps, weighs what frontier developers' safety frameworks are worth to a deployer, consolidates the organisational defences into six disciplines with a coverage matrix, learns what individuals, families, schools and libraries can do, and builds the AI threat register that holds the answer together, with entries, triggers, a one-page board report and a calibration log. Lecture: The Response: Law as Floor, Evidence as Standard, the Register as Memory

Module 7 · Part VII · The Verdict

Part VII delivers the verdict. The participant compares the answers Nordvik Region and Hansa Mutual gave to their governing bodies, weighs the optimistic visions and the five conditions for going well, and gathers the verdicts on all ten tropes into an answer to the title question, with the signals that would change it. They leave with a one-page threat statement, a positive-scenario entry and a forked register snapshot for their own organisation. Lecture: The Verdict: Something in It, but Not What the Stories Said

Final assignment

Capstone

An AI Threat Assessment and Statement for a Governing Body

Your governing body has asked whether the organisation should pause, restrict or continue its use of AI, or whether AI threat belongs in its risk assessment. Answer the question as it was asked, for your own organisation or for a third fictional organisation you define, using the method Nordvik Region and Hansa Mutual used. Write terms of reference, grade the claims behind the question, build an AI threat register that spans the four scales, map its controls and their evidence, and reduce your answer to one page. Everything must be traceable to evidence you can cite or to your organisation's own records; invent no statistics and assign no probability to speculative entries. Describe misuse and model behaviour at the level of evidence and governance only.

Steps

  1. Write terms of reference that frame the question exactly as asked, with scope across all four scales, team, external challenge, method and deliverables.
  2. Prepare three claim cards or evidence records for the claims that prompted the question, each challenged one rung up and one down, with dissent recorded.
  3. Build a register of at least eight entries covering every scale, each with tropes, rung, label, exposure, influence, controls, owner, decision tier and triggers in both directions.
  4. Produce two supporting artefacts from the course that close your largest gaps, for example a harm-path review, a stop-path test plan, a reversibility test or a callback-verification policy.
  5. Build a coverage matrix classifying each control as verified, declared or failed, with findings, owners and dates.
  6. Write the one-page threat statement and the one-page board report with at most three decisions, and present both to peers acting as the governing body.

Deliverables

  • Terms of reference and three graded, challenged claim cards or evidence records
  • An AI threat register of at least eight entries with a coverage matrix and findings
  • Two supporting artefacts chosen from the course
  • A one-page AI threat statement and a one-page board report

Assessment criteria

  • Every claim is stated precisely and graded at the highest rung its evidence reaches, with capability, propensity and realised harm kept apart
  • The register spans all four scales, separates exposure from influence, contains no invented probability and has triggers that can lower as well as raise concern
  • Controls are classified on evidence, and no control depends on the cooperation or self-report of the system it controls
  • The statement answers the question as asked, names what is declined and what remains unresolved, and keeps measures that rest on speculation reversible
  • Both the case for concern and the case for calm are represented at their strongest, and interpretation is marked as such