How to Design a Pre-Bunking Training Scenario for a Civic Audience

The Design Challenge

Pre-bunking training has a fundamental design tension: you need to expose participants to a manipulation technique convincingly enough that recognition is trained, but not so convincingly that the false content itself causes harm. Get the balance wrong in either direction and your training either fails to inoculate or becomes a vector for the very misinformation you're trying to counter.

The Talantir methodology, which underlies Retelnist's training module, resolves this tension through a specific scenario structure developed across deployments with national resilience programmes.

Step 1: Select the Technique, Not the Topic

The first and most important design decision: you are training recognition of a technique, not awareness of a specific narrative. Select the manipulation technique to train before you select any content.

The five techniques with the broadest transfer and most evidence-based training effects:

1. Emotional language amplification — using charged, fear-inducing, or outrage-triggering language to bypass analytical processing
2. False dichotomy construction — presenting complex situations as binary choices between only two options
3. Artificial social consensus — creating the impression that "everyone" or "most people" hold a position when the evidence does not support that claim
4. Identity threat framing — framing a policy or factual claim as an attack on in-group identity or values
5. Source credibility spoofing — appropriating the visual, stylistic, or attributional signals of trusted sources to transfer their credibility to false content

Step 2: Build the Scenario Structure

Each Talantir scenario has four components:

Component A: The Hook (30 seconds)

A realistic setup that places the participant in a plausible information consumption context. "You open Telegram and see the following post in a public channel you follow." The hook must feel authentic — if it feels obviously constructed, participants will not engage with it as if it were real.

Component B: The Example (60–90 seconds)

The actual manipulative content, constructed to use the target technique clearly enough to be recognisable but embedded in a realistic context. The content should be on a topic the audience is genuinely interested in — irrelevant topics produce disengaged participants and weak training effects.

Critical rule: the false factual claims in Component B must be visibly flagged as constructed training content before the scenario ends. Do not leave participants with unaddressed false claims.

Component C: The Reveal (45 seconds)

Explicit identification and naming of the technique used, with direct reference to the specific elements of Component B that demonstrate it. The naming is important — labelling the technique activates the conceptual anchor that enables future recognition.

Component D: The Transfer (60 seconds)

A different example of the same technique, applied to a different topic, that participants must identify themselves. Transfer practice is what converts recognition of a specific instance into generalised technique recognition. Without Component D, training effects are topic-specific rather than technique-general.

Step 3: Select the Training Content Topic

Now select the topic for your scenario. Effective topic selection criteria:

• Audience relevance: The topic should be one the audience genuinely encounters and cares about. Generic examples reduce engagement and training effectiveness.
• Moderate salience: Avoid topics with very high emotional charge for the target audience — the emotional response will overwhelm technique recognition. Avoid topics with no personal relevance — disengagement will prevent learning.
• Current in the information environment: Use topics where the technique is actually being deployed against the target population. This is where Retelnist's detection layer connects to the training layer — emerging narratives from Layer A monitoring become training scenario inputs.

Step 4: Write the Manipulative Content

This is the most delicate step. Writing genuinely persuasive manipulative content for training purposes requires skill and care.

Guidelines:

• The content should be realistic enough that a participant encountering it in the wild would not immediately dismiss it
• The target technique should be visible enough that a trained participant can identify it in the Reveal
• Include multiple realistic signals (consistent source styling, believable statistics that are actually fabricated, emotional language woven naturally) — over-obvious construction reduces training value
• All specific factual claims must either be true or be clearly labelled as constructed in the Reveal component

Step 5: Evaluate and Iterate

Pre-bunking training scenarios should be evaluated against measurable outcomes, not participant satisfaction scores. The relevant metrics:

• Technique recognition rate: In the Transfer component, what percentage of participants correctly identify the technique? Target: >75% for a well-designed scenario.
• Transfer distance: Test recognition on a third example not included in the training. Transfer distance measures how general the recognition has become.
• Retention at 30 days: Administer a recognition test 30 days after training. Strong scenarios show >60% retention without booster content.

Scenarios that do not meet these benchmarks should be redesigned or replaced. Talantir scenario library updates are informed by evaluation data from all active deployments, with poorly performing scenarios retired and replaced with evidence-based alternatives.

A Note on Sensitive Populations

For populations with recent trauma exposure (conflict zones, communities that have experienced targeted disinformation campaigns), standard scenario design requires modification. High emotional charge content, even for training purposes, can retraumatise. Consult with psychosocial support specialists before deploying training in these contexts.

Retelnist provides modified Talantir scenario sets for conflict-adjacent deployments on request.