Rubric · canonical · TECHNOLOGY-RISK-v1

Novel technology risk under uncertainty

For arguments about the risk profile of a novel technology where empirical history is limited and uncertainty is intrinsic.

Status canonical Criteria 4 (equally weighted) Updated May 2026

When this rubric applies

The argument is about whether a novel technology — an AI system, a biotech intervention, a geoengineering proposal — presents acceptable risk. The disagreement turns on how to reason about probability and consequence under genuine uncertainty, not on settled empirical findings.

Not appropriate when: the disagreement is about a settled empirical claim about the technology (use SCIENTIFIC-CLAIM) or about whether a benefit-cost calculation favors one outcome (use POLICY-TRADEOFF).

Criteria

1. Distinguishes intrinsic vs. deployment risks

weight 25%

Does the argument separate risks inherent in the technology from risks that depend on how it’s deployed?

10Distinguishes the two clearly; identifies which risks are intrinsic and which are deployment-dependent. 7Treats them as conceptually distinct but doesn’t apply the distinction systematically. 4Conflates them; argues against intrinsic risks using deployment fixes (or vice versa). 1Treats all risks as a single undifferentiated bucket.

2. Calibration of probability estimates

weight 25%

When the argument estimates probability, are the estimates calibrated, sourced, and explicit about uncertainty?

10Provides ranges with stated reasoning; acknowledges where the estimate is itself uncertain. 7Provides point estimates with implicit uncertainty. 4Asserts probabilities (high / low / negligible) without justification. 1Treats outcomes as certain when the evidence does not support that.

3. Treatment of irreversibility

weight 25%

Does the argument engage whether the harms (or benefits) are reversible? Irreversible outcomes warrant different decision rules than reversible ones.

10Engages reversibility explicitly; applies precautionary or expected-value reasoning appropriately. 7Acknowledges irreversibility for some risks; ignores it for others. 4Treats all outcomes as if they were equally reversible. 1Implicitly assumes reversibility while arguing for irreversible bets.

4. Engagement with mitigation pathways

weight 25%

Does the argument identify available mitigations — technical, regulatory, institutional — and assess their effectiveness?

10Names specific mitigation pathways and assesses their feasibility and likely effectiveness. 7Gestures at mitigations without specifics. 4Asserts that mitigation is or is not possible without engaging the actual options. 1Asserts technological inevitability or impossibility on either side.

Version history

v1May 2026 · initial publication