I. Authority Laundering Language
- “The model decided / the system decided”
Signals responsibility diffusion and misdelegation. - “AI-driven decision”
Collapses epistemic output into decision authority. - “Automated but human-supervised” (without naming authority)
Often masks delegation without remainder. - “Objective output”
Implies paradigm neutrality; usually false. - “Data-driven outcome”
Erases judgment, values, and decision ownership.
II. Legitimacy Substitution & Outcome Worship
- “It works well in practice” (used to bypass governance)
Treats performance as legitimacy. - “High confidence score” (used as authorization)
Confidence substituted for authority. - “Proven by benchmarks”
Benchmarks replacing contextual judgment. - “Best available option” (without contestability)
Used to foreclose alternatives prematurely. - “No better alternative”
Often a sign alternatives were never permitted.
III. Speculative Authority & Emergency Framing
- “Given the stakes, we can’t afford to wait”
Classic speculative authority capture. - “If we don’t act now…”
Time pressure used to bypass legitimacy. - “Existential risk requires exceptional measures”
Permanent emergency logic indicator. - “The future cost of inaction is too great”
Hypothetical futures overriding present authority. - “This is too important to slow down”
Anti-ACP framing; flags acceleration bias.
IV. Pseudo-Governance & Compliance Theater
- “We follow best practices” (unspecified)
Signals norm signaling without enforcement. - “Aligned with ethical principles”
Ethics invoked without procedural constraint. - “Reviewed internally” (no standing or appeal)
Often means unchallengeable. - “Responsible AI framework” (without refusal power)
Likely performative. - “Human-in-the-loop” (without authority transfer back)
The most common automation theater phrase.
V. Paradigm Smuggling & Neutrality Claims
- “Just the data”
Denies framing and value assumptions. - “Model-agnostic”
Often false; hides paradigm encoding. - “Politically neutral”
Signals unacknowledged value alignment. - “Purely technical decision”
Used to evade accountability. - “No subjective judgment involved”
Almost always incorrect in high-stakes contexts.
VI. Suppression of Contestability
- “We’ve already decided” (before review)
Indicates foreclosed contestation. - “That’s outside the scope of review” (used broadly)
Scope used to silence challenge. - “Let’s not reopen this”
Premature closure signal. - “Consensus has been reached” (without dissent record)
Consensus asserted, not demonstrated. - “Edge case” (used to dismiss legitimate concern)
Often a way to avoid paradigm stress-testing.
VII. Misframing of Refusal and Non-Action
- “ACP blocked progress”
Treats refusal as failure. - “Nothing happened” (after averted harm)
Failure-that-wasn’t erased. - “Overly cautious”
Used to delegitimize epistemic humility. - “Analysis paralysis”
Often means legitimate uncertainty was surfaced. - “We can fix it later”
Defers accountability downstream.
VIII. Capture via Rebranding & Drift
- “Agora-style” / “ACP-inspired”
Name dilution without constraint adoption. - “Lightweight ACP”
Often means ACP without refusal. - “ACP-compliant” (self-declared)
Compliance theater red flag. - “Governance-aware AI”
Awareness substituted for enforcement. - “Philosophy of ACP”
ACP reframed as worldview rather than protocol.
IX. Canonical ACP Warning Rule
If a phrase makes authority feel natural, inevitable, or ownerless, it is suspect.
If it makes refusal sound like failure, it is wrong.
X. How This Glossary Should Be Used
- As a lint tool for documents and PR
- As a review trigger in governance discussions
- As a training artifact for new collaborators
- As a regression test against ACP drift over time
Member discussion: