Determine the appropriate level of human involvement in AI decisions
Download: JSON · MarkdownUnder Singapore Model AI Governance Framework, Second edition, Part on human involvement in AI-augmented decision-making
What this label means
- Verified
- — a named reviewer opened the official source and confirmed the record; the date and the reviewer are shown.
- Pending review
- — linked to the official source and in the queue for a reviewer to confirm.
- Source-linked
- — linked to the official source, not yet confirmed by a reviewer.
On this page 9 sections
In brief
Determine the appropriate level of human involvement in AI decisions is a voluntary measure under the Singapore Model AI Governance Framework in Singapore (Second edition, Part on human involvement in AI-augmented decision-making). Using a risk-impact matrix (probability and severity of harm), organisations choose human-in-the-loop, human-over-the-loop or human-out-of-the-loop designs and document the rationale.
- Instrument
- Singapore Model AI Governance Framework
- Jurisdiction
- Singapore
- Nature
- Voluntary guidance
- Source reference
- Second edition, Part on human involvement in AI-augmented decision-making
- Who it binds
- Deployer / user organisation
- Evidence examples
- 1
- Framework mappings
- NIST AI Risk Management Framework 1.0
- Controls that meet it
- 2
- Official source
- Open official text
- Verification
- Source-linked
What does it require?
Using a risk-impact matrix (probability and severity of harm), organisations choose human-in-the-loop, human-over-the-loop or human-out-of-the-loop designs and document the rationale.
Practical action
Record the human-involvement choice and rationale for each AI decision point.
Who does it apply to?
Voluntary for any organisation deploying AI for decisions.
- Sectors
- Cross-sector / all sectors
Which controls meet this duty?
Satisfies: the control, operated properly, does the work the duty asks for. Supports: it contributes but the duty needs more. Each control page lists every other duty it serves, so work done once can be counted once.
-
satisfiesProcessAI system owner · once per ai systemHuman oversight design and override procedure
Serves 12 recorded duties · evidence: Human oversight and override procedure, Human-involvement design rationale, Overseer training completion
Records the in-the-loop, over-the-loop or out-of-the-loop choice and rationale.
-
supportsProcessAI system owner · once per ai systemAI risk assessment and lifecycle risk register
Serves 12 recorded duties · evidence: AI system risk assessment, Per-system AI risk register, Residual-risk acceptance
Probability and severity of harm drive the involvement level.
Templates that cover this duty
Generated from the records. Free · no account · links emailed to your work address. This duty is cited in each file with its source reference and a link back here.
- Human Oversight Procedure
Procedure · DOCX and XLSX · v5
- AI Agent Registry and Permission Matrix
Register · XLSX and DOCX · v6
- NIST AI RMF ↔ EU AI Act ↔ ISO/IEC 42001 Crosswalk
Crosswalk · XLSX · v3
- AI Audit Evidence Tracker
Register · XLSX · v1
What evidence would a reviewer expect?
| Evidence | Type | Notes |
|---|---|---|
| Risk-impact assessment and human-involvement decision | record |
Framework mappings
Original editorial crosswalks. They cite clause numbers only and reproduce no standard text; confidence reflects how direct the mapping is.
See every Singapore duty mapped this way →
| Framework | Reference | Note | Confidence |
|---|---|---|---|
| NIST AI RMF 1.0 | GOVERN 3.2, MANAGE 2.x | Original editorial mapping. | medium |
Frequently asked questions
Is this a legal requirement?
No. It is voluntary guidance under Singapore Model AI Governance Framework; following it is expected or recommended rather than required by law.
Who must comply with this obligation?
Voluntary for any organisation deploying AI for decisions.
What evidence shows compliance?
A reviewer would expect: Risk-impact assessment and human-involvement decision.
Which frameworks map to this obligation?
It is mapped to NIST AI Risk Management Framework 1.0 GOVERN 3.2, MANAGE 2.x. These are editorial crosswalks with a stated confidence, not official mappings.
Which controls satisfy this obligation?
AI risk assessment and lifecycle risk register; Human oversight design and override procedure.
What should an organisation do in practice?
Record the human-involvement choice and rationale for each AI decision point.
Cite this record
AIPolicyTracker (2026). “Determine the appropriate level of human involvement in AI decisions (Singapore Model AI Governance Framework)” [AIPT-OBL-0055, verified 26 September 2026 by Bhaskar Bhatt]. https://aipolicytracker.org/id/AIPT-OBL-0055 (accessed 27 October 2026). Data licensed CC BY 4.0.
Permanent id AIPT-OBL-0055: it never changes or gets reused, and its address keeps working if this page moves.
Cite the official text alongside it: Model AI Governance Framework (Second Edition), Personal Data Protection Commission Singapore and IMDA, https://www.pdpc.gov.sg/help-and-resources/2020/01/model-ai-governance-framework.
BibTeX
@misc{aipolicytracker2026_determinetheappropriatel,
author = {{AIPolicyTracker}},
title = {{Determine the appropriate level of human involvement in AI decisions (Singapore Model AI Governance Framework)}},
year = {2026},
howpublished = {\url{https://aipolicytracker.org/obligations/singapore-mgf-human-involvement}},
url = {https://aipolicytracker.org/obligations/singapore-mgf-human-involvement},
urldate = {2026-10-27},
note = {Data licensed CC BY 4.0. Accessed 27 October 2026}
}
Change history
No change to this record's tracked fields has been recorded since 15 October 2026, when field-level history began.
Similar obligations in other instruments
- Employers and employment agencies must let candidates request an alternative selection process or accommodation — NYC Local Law 144 (automated employment decision tools), New York (United States)
- Respect the right to object to automated decision-making without human intervention — UAE PDPL, United Arab Emirates
- Operators of high-impact AI must ensure human management and supervision — AI Basic Act, South Korea
- Apply safeguards to solely automated decisions with significant effects — ICO AI guidance, United Kingdom
- Enable and assign effective human oversight — EU AI Act, European Union
- Offer meaningful human review of an adverse consequential decision — Colorado ADMT law (SB 26-189), Colorado (United States)
- Provide routes to contest AI outcomes and seek redress — UK AI regulation framework, United Kingdom (voluntary)
- Test and monitor systems, enable human control, and be transparent with users (guardrails 4 to 6) — Australian Voluntary AI Safety Standard, Australia (voluntary)
Disclaimer: informational only, not legal advice. Verify every claim against the linked official sources and consult a qualified lawyer before acting.