---
id: "contrarian-values-vs-nightmares"
type: "contrarian-insight"
source_timestamps: ["§ Flaw #2: It doesn't provide a clear picture of what success looks like.", "§ Let's Talk About Nightmares"]
source_url: "https://hbr.org/2026/05/what-are-your-companys-ai-nightmares"
source_title: "What Are Your Company's AI Nightmares?"
tags: ["ethics", "framing", "psychology", "contrarian-insight"]
related: ["claim-values-wrong-start", "claim-nightmares-create-alignment"]
challenges: "The conventional view that AI governance must begin with a positive articulation of corporate ethical values."
sources: ["governance"]
sourceVaultSlug: "hbr-seg-governance"
originDay: 7
articleStem: "hbr-cl-82-ai-nightmares"
sourceUrl: "https://hbr.org/2026/05/what-are-your-companys-ai-nightmares"
sourceTitle: "What Are Your Company’s AI Nightmares?"
---
# Start with Nightmares, Not Values (Contrarian)

**The contrarian claim:** The conventional wisdom in AI ethics is to define positive organizational values (fairness, transparency, accountability) and build systems to uphold them. Blackman ([[entity-reid-blackman]]) argues this is **backwards**.

His reasoning: values are abstract, fail to define what success looks like (see [[claim-values-wrong-start]]), and don't motivate people. Instead, organizations should embrace a **"negative" framing** by explicitly defining their worst-case scenarios — nightmares. This leverages the *psychological urgency of disaster avoidance* to drive actual behavioral change and alignment (see [[claim-nightmares-create-alignment]] and the quote [[quote-lip-service-to-fairness]]). It is validated by the unnamed bank risk professional in [[quote-bank-risk-professional]].

**What it challenges:** The prevailing practice — used by major tech firms, standards bodies, and regulators — of beginning Responsible AI with a positive statement of principles and only then deriving controls.

**Enrichment note — the strongest counter-perspective:** Many practitioners argue values are still a *necessary* starting point because they (a) align AI efforts with mission, legal obligation, and societal norms, and (b) provide a common language for regulators and external stakeholders. On this view, starting from nightmares alone risks a patchwork of scenario-specific fixes without a coherent normative foundation, and values are needed to *prioritize* which nightmares matter most. The reconciling position: values and nightmares are **complementary, not mutually exclusive**. This tension surfaces directly in [[question-nightmare-disagreement]].


## Related across articles
- [[contrarian-total-safety-impossible]]
- [[action-ask-what-could-go-wrong]]
- [[contrarian-unanimous-support-warning]]