"They Are Supposed To": The Quiet Paradox Undermining Accountability

“They are supposed to…” might be the most corrosive five-word sentence fragment in modern governance—public or corporate. It sounds procedural, even rational. In reality it is often a linguistic escape hatch: a way to outsource responsibility upward, sideways, or into abstraction. The paradox is stark: at the very moment when governments, organizations, and companies loudly demand ever-greater transparency and accountability, everyday actors—managers, team leads, procurement officers, even regulators—deploy phrases that push true ownership away from themselves. The result: elegant frameworks on paper; brittle execution in practice. We have seen this pattern before in how organizations narrate AI responsibility—promising “ethical oversight” while leaving no one explicitly accountable (see our earlier piece on ethical AI gaps).

The Global Accountability Escalator

#

Across sectors, transparency mandates have multiplied: ESG disclosures, AI risk registers, supply chain traceability, data protection impact assessments. Each new requirement adds a rung to what should be an escalator lifting organizational maturity. Instead, diffusion language stalls the machinery. “Legal will handle it.” “Audit signs off.” “The platform team owns that.” No single abdication causes collapse, but the accumulation creates systemic drag—delays, blame loops, performative compliance—echoing warning signs described in our analysis of how ethics programs decay into messaging (The Ethics Alarm).

Asia’s Dual Mirror: High-Governance and Constraint Environments

#

Consider Singapore’s repeated emphasis on responsible data and AI deployment—reinforced by its Model AI Governance Framework (referenced globally for its clarity on explainability and human oversight). The framework presumes proactive cross-functional accountability: product teams adapting practices, leadership financing governance, engineers instrumenting traceability. Yet inside some firms, implementation meetings still stall on: “Security is supposed to provision that log pipeline.” Weeks pass. Progress reports show green boxes; underlying controls remain partial.

Shift to a different context: a mid-tier manufacturing supplier in a developing Asian economy facing stricter sustainability reporting from multinational buyers. Factory management echoes: “They are supposed to give us the emissions factors,” referring to an industry body; procurement counters: “They are supposed to train our staff.” In the vacuum, spreadsheets proliferate, numbers get approximated, and downstream brands unknowingly anchor decisions on soft data. The appearance of accountability masks fragile underlying processes.

The Linguistics of Responsibility Evasion

#

The phrase functions as a psychological deferral mechanism. By invoking an unnamed “they,” the speaker preserves personal reputational safety while projecting procedural alignment. Over time, this normalizes:

1. Ownership Dilution – Tasks become nouns without agents: “Compliance review,” “Governance sign-off”—status elements detached from named stewards.
2. Temporal Slippage – Without a named executor, deadlines become elastic; urgency diffuses.
3. Metric Theater – Dashboards celebrate volume (policies published, workshops held) while impact metrics (cycle time reduction, error detection lead time) go unowned.

The Reference Points—Used, Not Leaned On

#

Peter Drucker’s enduring principle—clarity of responsibility as a managerial cornerstone—remains timeless precisely because diffusion language attempts to dissolve that clarity. Likewise, Singapore’s AI governance articulation (focusing on explainability and internal accountability chains) exemplifies a modern regulatory attempt to force named ownership. These references are signals, not crutches: they show that our accountability dilemma is neither novel nor purely local.

Why Transparency Alone Fails

#

Transparency without proximate ownership is a hall of mirrors. Publishing logs, reports, or model cards does not create accountability unless a clearly empowered actor owns (as we argued when examining algorithmic decision contexts in AI life-impact decisions):

• Interpretation (What does this anomaly mean?)
• Escalation (Who is alerted? How fast?)
• Remediation (Who allocates budget and talent?)

In multiple post-incident reviews I’ve seen across industries—from infrastructure outages to ethics review backlogs—the root cause was not absence of data but absence of an accepted obligation to act on it.

Reframing the Paradox

#

We demand more accountability systems while perfecting accountability deflection. The phrase “they are supposed to” signals three failure modes:

Design Principles to Collapse the Escape Hatch

#

1. Mandate Named Stewards per Control – Every critical control (e.g., data lineage, model auditability, vendor risk rating) must list a primary and fallback owner—published internally.
2. Responsibility Retrospectives – After major initiatives, run a short review only on ownership clarity: Which tasks stalled due to implicit “they” assumptions? Archive and trend these.
3. Escalation Time SLAs – Track elapsed time from issue detection to acknowledgement by an owner. Use the metric as a leadership KPI—not just MTTR.
4. Accountability Language Hygiene – In status meetings, require translation: any “they are supposed to” must be followed by a named role and date. Non-compliant items are flagged amber.
5. Outcome-Centered Dashboards – Replace count metrics (policies published) with effectiveness metrics (policy-induced defect reduction, false positive decrease). Ownership becomes inseparable from performance.

Cultural Counterweights

#

Psychological safety is often misapplied as comfort with deferral. True safety enables early admission of gaps. Normalize statements like: “I have custody of this, and it is behind.” Precision beats polished ambiguity.

The Leadership Imperative

#

Executives must model refusal of faceless attribution. When leaders answer “They are supposed to” with “Name them or reassign to someone who can act,” they compress feedback loops and teach the organization that ambiguity is a cost center.

Closing Thought

#

Accountability is not a louder call for transparency. It is the disciplined assignment of obligation coupled with the courage to own imperfection publicly. Every time we permit “they are supposed to” to stand unchallenged, we trade short-term interpersonal ease for long-term systemic fragility. The paradox dissolves the moment we convert anonymous expectation into explicit stewardship. That conversion is a managerial act; neglecting it is a managerial failure. Building this cultural muscle complements the broader maturation of governance disciplines outlined in our overview of the rise of structured ethical practice (Moral Maze).

References (kept intentionally minimal):

1. Singapore Personal Data Protection Commission – Model AI Governance Framework (illustrative of explicit internal accountability emphasis): https://www.pdpc.gov.sg/Help-and-Resources/2020/01/Model-AI-Governance-Framework
2. Peter Drucker, “The Effective Executive” (classic management guidance on responsibility and effectiveness).