Why it works

Every design choice cites a source.

AI Safety 101 is built on peer-reviewed learning and behavioral science, not slide-deck habit. Below is the research each mechanic comes from - real papers, linked so you can check them. The citations describe general research; they are not evaluations of this course. The result feels like a modern product, not a compliance PDF.

Testing effect

Retrieval beats review

The finding: Taking a test on material produces far better long-term retention than re-studying it - even without feedback. In Roediger & Karpicke's experiments, tested learners dramatically outperformed re-studiers after a delay. Karpicke & Roediger then showed repeated retrieval is the critical factor in keeping knowledge accessible.

In the course: every module ends in a scored check, and sprint drills force recall under time pressure. Nobody passes by watching.

Roediger & Karpicke, Psychological Science (2006) →
Karpicke & Roediger, Science (2008) →
Scenario training

Decisions transfer to behavior

The finding: In Sitzmann's meta-analysis of computer-based simulation games, learners outperformed comparison groups on self-efficacy, declarative and procedural knowledge, and retention - with the largest gains when the simulation was embedded in instruction and learners could retry.

In the course: every module is a branching scenario built on a documented incident. Learners don't read about the $25M deepfake - they're on the call, deciding. Wrong choices are safe to make and retry here, which is exactly when the research says learning sticks.

Sitzmann, Personnel Psychology (2011) →
Generation effect

Making the call yourself

The finding: Information learners generate themselves is remembered substantially better than information they simply read - the "generation effect," established by Slamecka & Graf.

In the course: at each decision point the learner produces the judgment - wire it or verify, paste or don't - before seeing the consequence. The answer is never handed over first.

Slamecka & Graf, J. Experimental Psychology (1978) →
Spacing

Spread out or forgotten

The finding: Cepeda, Pashler, Vul, Wixted & Rohrer's quantitative synthesis of distributed-practice research: practice spread over time beats the same practice massed together, across tasks and populations.

In the course: the Monday Incident Brief, monthly new modules, and quarterly refreshers re-expose the same reflexes on a spacing schedule - the subscription structure is the pedagogy.

Cepeda et al., Psychological Bulletin (2006) →
Desirable difficulties

Harder now, stronger later

The finding: Bjork & Bjork's work on desirable difficulties: conditions that feel harder during training - time pressure, varied scenarios, retrieval over review - produce better long-term retention and transfer than smooth, easy practice.

In the course: decision points run under time pressure, incidents vary across industries, and the capstone drops the learner into an unfolding incident with the clock running. It feels harder than a slide deck because it works harder.

Bjork & Bjork, J. Applied Research in Memory & Cognition (2020) →
The behavioral scienceWhy short in-the-moment drills change behavior where annual slide decks don't.
The forgetting curve

Annual training decays in weeks

The finding: Murre & Dros' replication of Ebbinghaus' forgetting curve: memory for learned material drops steeply in the first days and weeks after a single exposure, then keeps sliding. A once-a-year course is mostly gone by the quarter's end.

In the course: the subscription structure is the countermeasure - weekly Brief, monthly new modules, refreshers on a spacing schedule. Re-exposure is built into the product, not left to chance.

Murre & Dros, PLOS ONE (2015) →
Implementation intentions

If-then plans beat good intentions

The finding: Gollwitzer's classic study of implementation intentions: people who rehearse a specific "if situation X, then I do Y" plan follow through dramatically more reliably than people who merely intend to do the right thing.

In the course: every decision point is that rehearsal. If the CFO calls with urgency, then I verify out-of-band. If a document is sensitive, then it never touches a prompt. Learners practice the if-then until it's loaded.

Gollwitzer, American Psychologist (1999) →
Habit formation

Repetition builds the reflex

The finding: Lally and colleagues tracked real-world habit formation: a behavior becomes automatic through consistent repetition in a consistent context - on average over weeks, not from a single exposure, however vivid.

In the course: sprint drills repeat the same verification move under time pressure, and fresh modules keep the repetitions coming. The reflex is the product.

Lally et al., European J. Social Psychology (2010) →
COM-B model

Capability, opportunity, motivation

The finding: Michie, van Stralen & West's behaviour-change wheel: behavior changes when capability, opportunity, and motivation all move together. Awareness campaigns move motivation alone - and fail.

In the course: drills build capability, the policy template shapes the environment people work in, and real incidents supply the motivation - this actually happened, to a company like yours.

Michie et al., Implementation Science (2011) →
Behavior model

Small, prompted, in the moment

The finding: Fogg's behavior model: a behavior happens when motivation, ability, and a prompt converge in the same moment. Make the action small and prompt it at the right time, and it gets done; make it big, and it gets deferred.

In the course: modules are short on purpose - done in a coffee break, not deferred to next quarter - and the Monday Brief is the weekly prompt that brings people back.

Fogg, Persuasive Technology (2009) →
These citations describe general learning and behavioral research, not studies of this course. What we claim: the course is built the way this research says training should be built - and every completion produces the evidence file to show for it.