2026 04 26 implicit rate limiting controls agentic ai removal

Implicit rate-limiting controls removed by agentic Artificial Intelligence (AI): blast radius amplification and the operational risk literature gap

Research Question

Prior to agentic Artificial Intelligence (AI), the blast radius of ungoverned citizen development was implicitly bounded by human speed, attention, fatigue, and working hours, controls that are not documented in any risk framework but function as real operational constraints. Agentic AI removes these implicit rate-limiting controls without replacing them with engineered controls. Does any operational risk framework, automation risk framework, or AI governance framework explicitly account for the removal of these implicit controls, or does this represent a gap in current frameworks that must be constructed from the operational risk literature on automation and speed of consequence?

Scope

In scope:

The concept of implicit controls: undocumented constraints that function as real risk mitigants because they arise from human operational characteristics (speed, attention, fatigue, working hours, cognitive load)
Evidence that human speed, attention, fatigue, and working hours have historically bounded the blast radius of ungoverned citizen development and automation
Whether any operational risk framework, including Basel Committee principles, Australian Prudential Regulation Authority (APRA) CPS 230, Digital Operational Resilience Act (DORA), International Organization for Standardization (ISO) 31000, National Institute of Standards and Technology (NIST) Special Publication (SP) 800-53, or the operational risk literature more broadly, explicitly accounts for implicit controls as a risk mitigant category
Whether any AI governance framework or automation risk framework explicitly accounts for the removal of implicit rate-limiting controls as a transition risk when moving from human to agentic operation
The operational risk literature on automation failure, speed of consequence, and normal accident theory, as a potential first-principles foundation if framework-level coverage is absent
Analogues from other technology transitions that removed implicit human rate-limiting: high-frequency trading (HFT), robotic process automation (RPA), algorithmic trading, automated clearing and settlement

Out of scope:

General automation risk management beyond the specific implicit-controls question
The access control amplification problem (covered by the companion amplification item)
Technical controls engineering to replace implicit controls (this item establishes the gap; controls design is a separate concern)
AI alignment or model behaviour under adversarial inputs

Constraints:

The primary question is whether implicit controls are named and accounted for in frameworks, not whether the underlying risk is acknowledged generally
Evidence from financial services technology analogues (HFT, RPA, algorithmic trading) is directly applicable
Distinguish between frameworks that acknowledge automation speed as a risk factor versus frameworks that explicitly name the removal of implicit human controls as a mechanism requiring engineered replacement

Context

[fact; source: https://davidamitchell.github.io/Research/research/2026-04-26-systems-capability-debt-agentic-ai-risk-synthesis.html; https://davidamitchell.github.io/Research/research/2026-04-26-access-control-amplification-agentic-operations.html; https://davidamitchell.github.io/Research/research/2026-04-26-agentic-ai-regulatory-preconditions-control-failure-assessment.html] Prior completed repository work already established the broader systems-capability-debt to citizen-development to agentic-risk chain, the adjacent access-control amplification mechanism, and the regulatory argument that unresolved foundational control weakness is already a control-failure issue in regulated settings.
[inference; source: https://davidamitchell.github.io/Research/research/2026-04-26-systems-capability-debt-agentic-ai-risk-synthesis.html; https://davidamitchell.github.io/Research/research/2026-04-26-access-control-amplification-agentic-operations.html] This companion item isolates a narrower mechanism inside that chain: when work moves from human-paced execution to continuous autonomous execution, undocumented human operating constraints disappear and blast radius increases even if formal permissions and business processes look unchanged on paper.

Approach

Define implicit controls conceptually: Draw on the operational risk and systems safety literature to define implicit controls, undocumented constraints that function as real mitigants, and distinguish them from formal, designed controls. Assess whether this concept has a recognised name in the literature (e.g., residual friction, natural barriers, soft controls).
Evidence that human factors have bounded citizen development blast radius: Review the citizen development and shadow IT (Shadow Information Technology) literature for evidence, direct or indirect, that human speed and attention have historically limited the operational damage from ungoverned automation. The RPA literature may be the most proximate source.
Framework survey for explicit implicit-controls language: For each of Basel Committee operational resilience principles, APRA CPS 230, DORA, ISO 31000, NIST SP 800-53, NIST Artificial Intelligence Risk Management Framework (AI RMF) 1.0, and the broader operational risk literature, search for provisions that explicitly account for implicit controls as a risk mitigant or for the removal of such controls as a deployment transition risk.
Technology analogue review: Review the HFT, algorithmic trading, and RPA literature for evidence of how those transitions handled, or failed to handle, the removal of implicit human rate-limiting controls. The 2010 Flash Crash and RPA incident literature are candidate sources.
Gap characterisation: If no framework explicitly addresses the removal of implicit controls, characterise the gap precisely: what would a framework provision addressing this mechanism need to say, and where would it sit in operational risk, AI governance, or systems safety frameworks?
Synthesis: Produce a clear statement of whether the implicit-controls removal mechanism is framework-covered, first-principles, or a genuine gap, and what the evidence base supports for a board risk committee briefing.

Sources

Research Skill Output

(Full output from running the research skill - retained verbatim in the completed item. Sections 0-5 are the investigation, and section 6 seeds the Findings section below.)

§0 Initialise

[fact; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.iso.org/standard/65694.html] Research question restated: do existing operational-risk, resilience, and AI-governance frameworks explicitly recognise that human speed, attention, fatigue, and working hours have functioned as de facto controls, and that moving work to continuous agentic execution removes those controls unless engineered replacements are introduced?
[fact; source: https://davidamitchell.github.io/Research/research/2026-04-26-systems-capability-debt-agentic-ai-risk-synthesis.html; https://davidamitchell.github.io/Research/research/2026-04-26-access-control-amplification-agentic-operations.html; https://davidamitchell.github.io/Research/research/2026-04-26-agentic-ai-regulatory-preconditions-control-failure-assessment.html] Prior completed repository work already covered the broader causal chain, access-control amplification, and regulatory preconditions, so this item tests a narrower mechanism inside that synthesis rather than re-proving the whole chain.
[fact; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] Scope confirmed: the test is not whether frameworks recognise automation risk generally, but whether they explicitly name implicit human operating constraints as a control category or explicitly require control substitution when autonomy removes those constraints.
[fact; source: https://academic.oup.com/rfs/article/32/5/2024/5428080; https://www.gartner.com/en/documents/3983454; https://www.sec.gov/files/marketevents-report.pdf; https://www.iso.org/standard/65694.html] Constraints confirmed: several seeded sources were inaccessible or incomplete in this runtime, so primary claims rely on accessible regulator texts, the official SEC and CFTC report, accessible academic RPA literature, and official framework summaries, with inaccessible sources recorded as access notes rather than used as support.
[fact; source: https://davidamitchell.github.io/Research/research/2026-04-26-systems-capability-debt-agentic-ai-risk-synthesis.html] Output format: knowledge.

§1 Question Decomposition

Root question: Is the removal of implicit human rate-limiting explicitly covered by current frameworks, or is it a first-principles gap?
A. Concept definition
- A1. What qualifies as an implicit control rather than a formal designed control?
- A2. Does the literature give this concept an established name?
B. Historical bounding
- B1. What evidence shows humans were acting as the glue, throttle, or exception handler in workaround-heavy estates?
- B2. What changes when those tasks move to continuous automation?
C. Framework survey
- C1. What do Basel Committee operational-risk and resilience principles explicitly require?
- C2. What does APRA CPS 230 explicitly require?
- C3. What does DORA explicitly require?
- C4. What do NIST AI RMF 1.0, NIST SP 800-53, and ISO 31000 explicitly require at public-text level?
- C5. Do any of those texts explicitly call human speed, fatigue, attention, or working hours controls?
D. Technology analogues
- D1. What does the Flash Crash record show about consequence speed and engineered pauses?
- D2. What does the RPA literature show about exception routing, monitoring, and unnoticed failure?
E. Gap characterization
- E1. If explicit framework coverage is absent, what first-principles explanation remains strongest?
- E2. What would a framework clause covering this mechanism need to require?

§2 Investigation

Search-record summary

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] Queries such as "implicit controls" operational risk framework, "removal of implicit human controls" automation risk, and "human speed" control framework did not locate reviewed framework text that explicitly names removed human speed, attention, fatigue, or working-hours limits as controls.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html; https://doi.org/10.1007/s12599-018-0542-4; https://link.springer.com/article/10.1007/s10257-022-00553-8; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists; https://press.princeton.edu/books/paperback/9780691004129/normal-accidents] Queries such as "residual friction" operational risk automation and "natural barriers" operational risk automation did not locate a settled operational-risk taxonomy in the reviewed framework and analogue literature that treats residual human friction or natural barriers as a formal control class.

A. Conceptual status of implicit controls

[fact; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents] Charles Perrow's normal accident theory explains risk through interactive complexity and tight versus loose coupling, and argues that adding safeguards can also increase complexity and create new categories of accidents.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents] Perrow provides the closest first-principles account of why implicit human friction matters, because slack, delay, and interpretive intervention are more available in loosely coupled human workflows than in tightly coupled automated ones, but he does not name those factors "implicit controls."
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] Across the reviewed framework texts and official summaries, I found extensive language on explicit controls, monitoring, change management, continuity, and oversight, but no text that explicitly classifies human speed, attention, fatigue, or working hours as control objects.

B. Evidence that human factors historically bounded blast radius

[fact; source: https://doi.org/10.1007/s10257-020-00472-6] The shadow-information-technology literature records business units sourcing or adapting their own systems when the central information technology function could not provide a suitable solution quickly enough, which shows workaround estates arise inside delivery-speed gaps rather than only from tool enthusiasm.
[fact; source: https://doi.org/10.1007/s12599-018-0542-4] The RPA literature states that humans often serve as the glue between disconnected information systems, manually entering information repeatedly and making decisions in the long tail of work that was not economical to automate previously.
[fact; source: https://doi.org/10.1007/s12599-018-0542-4] The same RPA source states that when a case turns out to be exceptional, the RPA agent may hand over the case to a human, and it warns that RPA agents can start making incorrect decisions because of contextual changes that may remain unnoticed for some time and lead to disastrous situations.
[fact; source: https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] ISACA describes RPA command-and-control centers that initiate, reinitiate, maintain, and monitor continuous, scheduled, or on-demand routines, and it identifies privileged access, hard-coded credentials, insufficient monitoring, and weak change management as material risk areas.
[inference; source: https://doi.org/10.1007/s12599-018-0542-4; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] Taken together, these sources show that pre-automation human work had been providing pace, selective attention, and bounded operating windows as de facto friction, and that once automation widened operating hours and execution speed, organizations needed explicit orchestration, exception handling, monitoring, and access controls to replace that friction.
[assumption; source: https://doi.org/10.1007/s12599-018-0542-4; https://doi.org/10.1007/s10257-020-00472-6] No directly observed study in the reviewed corpus measures "human speed" or "working hours" as a quantified blast-radius cap for citizen development specifically, so the claim that those factors historically bounded damage is inferred from shadow-information-technology and automation analogues rather than from a direct pre-agentic measurement study.

C. Framework survey

[fact; source: https://www.bis.org/fsi/fsisummaries/psmor.htm] Basel Committee operational-risk principles require comprehensive identification and assessment of operational risk in products, activities, processes, and systems, explicit change-management oversight controls, strong control environments with segregation of duties, monitoring and reporting, and robust Information and Communication Technology (ICT) risk management.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://www.bis.org/bcbs/publ/d516.htm] Basel addresses explicit control design and resilience obligations, but I found no Basel text that explicitly says human speed, attention, fatigue, or working hours were previously functioning as controls whose removal must trigger engineered replacement.
[fact; source: https://handbook.apra.gov.au/standard/cps-230] APRA CPS 230 requires entities to identify, assess, and manage operational risk with effective internal controls, monitoring, and remediation, maintain critical operations through severe disruptions, and manage service-provider risk.
[inference; source: https://handbook.apra.gov.au/standard/cps-230] CPS 230 clearly covers explicit operational-risk controls, but I found no provision that explicitly categorises human operating constraints as controls or names their removal during automation as a separate transition-risk mechanism.
[fact; source: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554] DORA Article 5 requires an internal governance and control framework for effective and prudent ICT-risk management, Article 6 requires a sound and comprehensive ICT-risk-management framework that is reviewed and continuously improved, Article 9 requires continuous monitoring and control of ICT systems plus automated isolation, access-rights controls, and change-management controls, Article 10 requires multiple layers of control and automatic alert mechanisms, and Article 11 requires business impact analysis, continuity planning, and interdependency-aware recovery arrangements.
[inference; source: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554] DORA comes closest to the implicit-controls issue because it explicitly requires automated containment and continuous monitoring in interconnected environments, but it still frames those as designed ICT controls and does not explicitly state that they are replacing removed human pace or working-hour constraints.
[fact; source: https://airc.nist.gov/airmf-resources/airmf/5-sec-core/] NIST AI RMF 1.0 requires continuous and lifecycle-wide risk management, roles and responsibilities for human-AI configurations and oversight, ongoing monitoring and periodic review, inventorying systems, and decommissioning or phasing out systems safely.
[inference; source: https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-ai-rmf-10] NIST AI RMF explicitly recognises that human oversight must be designed rather than assumed, but I found no NIST AI RMF text that explicitly instructs organizations to identify which human operating constraints disappear when an activity moves from human execution to continuous agentic execution.
[fact; source: https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final] The public NIST SP 800-53 publication states that the catalog addresses human errors as part of organization-wide risk management and includes control families for Access Control, Assessment, Authorization and Monitoring, Configuration Management, Contingency Planning, Risk Assessment, and System and Information Integrity.
[inference; source: https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final] At the public-summary level accessible in this runtime, SP 800-53 clearly assumes explicit control design, monitoring, and configuration control, but it does not publicly summarise the removal of human rate limits as a named risk mechanism.
[fact; source: https://www.iso.org/standard/65694.html] ISO 31000 says risk management should be integrated into governance, strategy, planning, reporting, processes, and culture, and should include identifying, analyzing, evaluating, treating, monitoring, communicating, and continually improving risk management.
[inference; source: https://www.iso.org/standard/65694.html] ISO 31000 provides a broad process discipline for identifying changed risk conditions, but the public summary does not explicitly discuss automation speed, implicit human controls, or control substitution when autonomy increases.

D. Technology analogues on speed of consequence

[fact; source: https://www.sec.gov/files/marketevents-report.pdf] The SEC and CFTC Flash Crash report says the large trader's May 6 sell algorithm targeted trading volume rather than price or time and executed the sell program extremely rapidly in just 20 minutes, whereas a prior comparable sell program had taken more than five hours to execute the first 75,000 contracts.
[fact; source: https://www.sec.gov/files/marketevents-report.pdf] The same report says many liquidity providers temporarily paused their automated trading systems during the liquidity crisis because built-in pauses were designed to let traders and risk managers fully assess market conditions before trading resumed.
[fact; source: https://www.sec.gov/files/marketevents-report.pdf] The report also says pausing a market can provide time for participants to reassess strategies and for algorithms to reset parameters, and it records the post-event circuit-breaker program that pauses trading in individual securities for five minutes after a 10% price change in five minutes.
[inference; source: https://www.sec.gov/files/marketevents-report.pdf] The Flash Crash record does not use the term "implicit controls," but it is strong evidence that once event speed outran ordinary assessment windows, markets had to recreate time for human and system reassessment through explicit engineered brakes.
[fact; source: https://doi.org/10.1007/s12599-018-0542-4; https://link.springer.com/article/10.1007/s10257-022-00553-8] The academic RPA literature describes RPA as replacing humans in manual long-tail work across existing interfaces, stresses that expert knowledge is required to make robots reliable and scalable, and reports estimates that up to 50% of initial RPA implementations fail.
[fact; source: https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] The practitioner RPA literature responds to that risk profile with command-and-control centers, continuous monitoring, audit trails, change management, and privileged-access governance, which are all explicit engineered controls around automation that is no longer bounded by ordinary human pace.

E. Gap-characterisation inputs

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.iso.org/standard/65694.html; https://www.sec.gov/files/marketevents-report.pdf] The evidence supports a consistent pattern: frameworks govern explicit controls, while the analogues show that when automation removes human pacing, organizations later add explicit rate limits, pauses, exception routing, segmentation, and monitoring, which implies an unfilled conceptual space between the two.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4] The strongest first-principles explanation for that space is that human-operated systems often contain slack and selective intervention by default, whereas tighter coupling and higher execution speed force organizations to build explicit interruption and review mechanisms once that slack disappears.

§3 Reasoning

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] The reviewed frameworks clearly recognise the need for explicit controls, monitoring, change management, resilience, and oversight, so the question is not whether control discipline exists but whether implicit human constraints are named as part of that discipline.
[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] The analogue evidence shows that once automation compresses time or extends operating windows, organizations respond by adding engineered pauses, monitoring, exception routing, and access controls, which is what one would expect if removed human pacing had previously been providing risk mitigation.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents] Perrow's coupling argument explains why the removal of slack matters: faster and more tightly coupled execution shortens the time available for interpretation, challenge, and interruption, so consequence propagation speeds up unless new brakes are designed.
[inference; source: https://doi.org/10.1007/s10257-020-00472-6; https://doi.org/10.1007/s12599-018-0542-4] Because workaround estates arise where central delivery has failed to meet local need and because humans were often manually stitching systems together, agentic execution over the same estate is not a neutral productivity shift but a transition that changes the control surface.

§4 Consistency Check

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.iso.org/standard/65694.html] No reviewed framework text contradicted the conclusion that explicit control, monitoring, and resilience obligations already exist, and no reviewed framework text explicitly named the removal of human pace, fatigue, attention, or working hours as a control-substitution issue.
[inference; source: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/] The strongest adjacent coverage came from DORA and NIST AI RMF, but both still require inference to connect their explicit control obligations to the narrower mechanism tested in this item.
[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4] The analogue evidence is consistent with, rather than contrary to, the framework survey: where speed and continuity increased, explicit brakes and exception paths had to be introduced.

§5 Depth and Breadth Expansion

[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554] Technical lens: the mechanism is best understood as a coupling problem, because autonomous execution shortens intervention windows and increases interdependency speed, which is why DORA emphasises continuous monitoring, automatic isolation, and interdependency-aware continuity planning.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://www.iso.org/standard/65694.html] Regulatory lens: prudential and risk-management frameworks are strong enough to criticise unmanaged autonomy under existing explicit-control duties, but they do not provide the precise explanatory language boards and committees need when the risk increase comes from removal of undeclared human friction.
[inference; source: https://doi.org/10.1007/s10257-020-00472-6; https://doi.org/10.1007/s12599-018-0542-4] Economic and organizational lens: workaround estates emerge when central delivery cannot meet demand fast enough, and automation's attraction is often to absorb manual long-tail work cheaply, so the pressure to deploy agents is structurally strongest exactly where latent control substitution is most likely to be missed.
[inference; source: https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists; https://www.sec.gov/files/marketevents-report.pdf] Behavioral lens: fatigue, working hours, and attention are poor formal controls but real operational constraints, and the analogue record shows organizations repeatedly convert those informal constraints into engineered pauses, review points, and monitoring only after automation exposes the missing design.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/] A framework clause that closed the gap would need to require organizations to inventory human-derived friction before autonomy increases, assess the blast-radius delta caused by its removal, and prove that rate limits, approval thresholds, segmentation, monitoring, exception routing, and shutoff mechanisms replace the lost mitigation.

§6 Synthesis

(This section seeds the Findings below.)

Executive summary:

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] The reviewed operational-risk, resilience, and AI-governance frameworks do not explicitly account for the removal of human speed, attention, fatigue, or working-hour constraints as a named control-substitution problem; that mechanism is a genuine conceptual gap in explicit framework language.
[fact; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] Those frameworks do, however, require explicit governance, monitoring, change control, resilience planning, and risk review, so unmanaged autonomy can still be criticised under existing duties even though the narrower mechanism is not named.
[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] The analogue evidence from high-frequency trading and RPA shows that once automation compresses time or extends operating windows, organizations add engineered pauses, monitoring, exception routing, and access controls to recreate the friction humans had been supplying implicitly.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://www.sec.gov/files/marketevents-report.pdf] Perrow's normal accident theory is the strongest first-principles foundation for the missing mechanism because tighter coupling and faster consequence propagation explain why blast radius increases when human slack disappears.

Key findings:

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] Confidence: high. The reviewed frameworks require explicit governance, monitoring, change control, resilience, and risk-management processes, but none of the reviewed texts explicitly classify human speed, attention, fatigue, or working hours as controls whose removal must be replaced with engineered controls.
[inference; source: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554] Confidence: medium. DORA comes closest to the missing mechanism because it mandates continuous monitoring, automatic alerting, automated isolation, controlled change management, and interdependency-aware continuity planning, yet it still frames those as explicit ICT controls rather than as substitutes for removed human rate limits.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230] Confidence: high. Basel Committee operational-risk principles and APRA CPS 230 clearly require strong control environments, oversight controls for change, monitoring, remediation, and ICT risk management, but they do not name pre-existing human operational friction as a distinct mitigant category.
[inference; source: https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-ai-rmf-10] Confidence: medium. NIST AI RMF 1.0 explicitly requires documented roles for human-AI configurations and oversight, ongoing monitoring, risk-tolerance setting, and safe decommissioning, but it does not instruct organizations to identify which human operating characteristics disappear when work shifts to continuous autonomous execution.
[fact; source: https://www.sec.gov/files/marketevents-report.pdf] Confidence: medium. The SEC and CFTC Flash Crash report shows that a volume-driven automated sell program compressed execution into 20 minutes, many liquidity providers paused simultaneously to reassess conditions, and exchanges then implemented circuit breakers to recreate assessment time explicitly.
[inference; source: https://doi.org/10.1007/s12599-018-0542-4; https://link.springer.com/article/10.1007/s10257-022-00553-8; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] Confidence: high. The RPA literature shows that humans had been serving as the glue across disconnected systems, that automation moves more of that long-tail work into software agents, and that reliable scaling then depends on explicit exception handling, monitoring, privileged-access control, and change management.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4] Confidence: medium. Perrow's normal accident theory provides the clearest first-principles explanation for the gap because tighter coupling and interactive complexity make consequence propagation faster and harder to interrupt once human slack is removed.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.iso.org/standard/65694.html] Confidence: medium. A framework provision that closed the gap would need to require organizations to inventory human-derived friction before autonomy increases and to prove that rate limits, approval thresholds, segmentation, monitoring, exception routing, and shutoff mechanisms replace the lost mitigation.

Evidence map:

Claim	Source	Confidence	Notes
[inference] No reviewed framework explicitly names human speed, attention, fatigue, or working hours as controls whose removal must trigger engineered replacement.	https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html	high	Absence claim limited to reviewed accessible texts and public summaries.
[inference] DORA is the closest adjacent framework because it requires continuous monitoring, automatic alerting, automated isolation, controlled change management, and interdependency-aware continuity planning without explicitly naming removed human rate limits.	https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554	medium	Strongest control-substitution analogue in reviewed framework set.
[inference] Basel Committee principles and APRA CPS 230 require explicit control environments, monitoring, remediation, and ICT risk management, but not the narrower implicit-controls mechanism.	https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230	high	Strong prudential support for critique, weak explicit naming of mechanism.
[inference] NIST AI RMF 1.0 requires designed human-AI oversight, monitoring, and safe decommissioning, but does not explicitly require identification of removed human operating constraints.	https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-ai-rmf-10	medium	Shows explicit oversight design without the narrower concept.
[fact] The Flash Crash record documents automated execution compressed to 20 minutes, liquidity-provider pauses, and post-event circuit breakers that recreated assessment time explicitly.	https://www.sec.gov/files/marketevents-report.pdf	medium	Primary analogue for speed-of-consequence and engineered brakes.
[inference] RPA evidence shows humans were previously acting as the glue across systems and that scaled automation requires explicit exception handling, monitoring, privileged-access control, and change management.	https://doi.org/10.1007/s12599-018-0542-4; https://link.springer.com/article/10.1007/s10257-022-00553-8; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists	high	Combines academic and practitioner evidence.
[inference] Perrow's normal accident theory best explains why removal of human slack increases blast radius in tightly coupled systems.	https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://www.sec.gov/files/marketevents-report.pdf	medium	Conceptual support rather than a framework citation.
[inference] A gap-closing framework clause would need to require inventory and replacement of human-derived friction before autonomy scales.	https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.iso.org/standard/65694.html	medium	Derived from reviewed duties plus analogue evidence.

Assumptions:

[assumption; source: https://doi.org/10.1007/s12599-018-0542-4; https://doi.org/10.1007/s10257-020-00472-6] The claim that human speed, attention, fatigue, and working hours historically bounded citizen-development blast radius is inferred indirectly from shadow-information-technology and automation analogues because the reviewed corpus did not contain a direct pre-agentic measurement study of that specific causal mechanism.
[assumption; source: https://academic.oup.com/rfs/article/32/5/2024/5428080; https://www.gartner.com/en/documents/3983454; https://www.iso.org/standard/65694.html; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final] The explicit-gap conclusion is limited to the reviewed accessible framework texts and public summaries, so a paywalled or inaccessible source could contain adjacent wording not visible in this runtime, although no evidence found here suggests that it does.

Analysis:

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] I weighted primary framework texts and official summaries more heavily than vendor or practitioner commentary when determining whether explicit coverage exists, so the explicit-gap conclusion is driven by regulator and standards documents rather than by later commentary.
[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] I used analogue evidence from high-frequency trading and RPA not to prove that the frameworks already contain the mechanism, but to test whether consequence speed, exception routing, and control substitution behave in practice the way the implicit-controls hypothesis predicts.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554] Perrow's coupling model and DORA's explicit interdependency and continuity language align strongly enough that the first-principles argument is more credible than a pure novelty claim detached from operational-risk theory.
[inference; source: https://doi.org/10.1007/s10257-020-00472-6; https://doi.org/10.1007/s12599-018-0542-4] The most important trade-off in the evidence is between directness and relevance: the shadow-information-technology and RPA sources are highly relevant to workaround estates but indirect on the exact phrase "implicit controls," while the framework texts are direct on explicit controls but silent on the narrower mechanism.

Risks, gaps, uncertainties:

[assumption; source: https://doi.org/10.1007/s12599-018-0542-4; https://doi.org/10.1007/s10257-020-00472-6] Direct empirical studies that quantify human pace, attention, fatigue, or working hours as a measured blast-radius cap for citizen development were not found in the reviewed corpus.
[assumption; source: https://academic.oup.com/rfs/article/32/5/2024/5428080; https://www.gartner.com/en/documents/3983454] The inaccessible Oxford and Gartner sources may contain additional supporting detail for the technology-analogue section, but they do not change the reviewed-framework conclusion because the explicit-gap claim is grounded in accessible primary framework texts.
[assumption; source: https://www.iso.org/standard/65694.html; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final] ISO 31000 and NIST SP 800-53 were only accessible at public-summary level in this runtime, so claims about their silence on the mechanism are lower precision than the DORA, Basel, APRA, NIST AI RMF, SEC, and RPA claims.

Open questions:

[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4] Which concrete control patterns best replace lost human friction in enterprise agent deployments, for example rate limits, mandatory approval thresholds, segmentation, exception-routing thresholds, or time-bounded credentials?
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/] How should boards and risk committees quantify blast-radius change when a workflow moves from human execution to continuous agent execution under otherwise unchanged permissions and process maps?

§7 Recursive Review

[fact; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html; https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4; https://link.springer.com/article/10.1007/s10257-022-00553-8; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists; https://press.princeton.edu/books/paperback/9780691004129/normal-accidents] Every factual claim in Sections 0, 2, and 6 is bound to an accessible source or clearly marked as an assumption or inference.
[fact; source: https://academic.oup.com/rfs/article/32/5/2024/5428080; https://www.gartner.com/en/documents/3983454] Inaccessible seeded sources were recorded as access notes and were not used as support for downstream factual claims.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4] The conclusion remained stable after re-checking the framework survey against the analogue evidence: explicit adjacent control obligations exist, but the narrower implicit-controls-removal mechanism remains first-principles rather than framework-explicit.

Findings

(Populated from Section 6 Synthesis above.)

Executive Summary

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] The reviewed operational-risk, resilience, and AI-governance frameworks do not explicitly account for the removal of human speed, attention, fatigue, or working-hour constraints as a named control-substitution problem; that mechanism is a genuine conceptual gap in explicit framework language.
[fact; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] Those frameworks do, however, require explicit governance, monitoring, change control, resilience planning, and risk review, so unmanaged autonomy can still be criticised under existing duties even though the narrower mechanism is not named.
[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] The analogue evidence from high-frequency trading and RPA shows that once automation compresses time or extends operating windows, organizations add engineered pauses, monitoring, exception routing, and access controls to recreate the friction humans had been supplying implicitly.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://www.sec.gov/files/marketevents-report.pdf] Perrow's normal accident theory is the strongest first-principles foundation for the missing mechanism because tighter coupling and faster consequence propagation explain why blast radius increases when human slack disappears.

Key Findings

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] Confidence: high. The reviewed frameworks require explicit governance, monitoring, change control, resilience, and risk-management processes, but none of the reviewed texts explicitly classify human speed, attention, fatigue, or working hours as controls whose removal must be replaced with engineered controls.
[inference; source: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554] Confidence: medium. DORA comes closest to the missing mechanism because it mandates continuous monitoring, automatic alerting, automated isolation, controlled change management, and interdependency-aware continuity planning, yet it still frames those as explicit ICT controls rather than as substitutes for removed human rate limits.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230] Confidence: high. Basel Committee operational-risk principles and APRA CPS 230 clearly require strong control environments, oversight controls for change, monitoring, remediation, and ICT risk management, but they do not name pre-existing human operational friction as a distinct mitigant category.
[inference; source: https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-ai-rmf-10] Confidence: medium. NIST AI RMF 1.0 explicitly requires documented roles for human-AI configurations and oversight, ongoing monitoring, risk-tolerance setting, and safe decommissioning, but it does not instruct organizations to identify which human operating characteristics disappear when work shifts to continuous autonomous execution.
[fact; source: https://www.sec.gov/files/marketevents-report.pdf] Confidence: medium. The SEC and CFTC Flash Crash report shows that a volume-driven automated sell program compressed execution into 20 minutes, many liquidity providers paused simultaneously to reassess conditions, and exchanges then implemented circuit breakers to recreate assessment time explicitly.
[inference; source: https://doi.org/10.1007/s12599-018-0542-4; https://link.springer.com/article/10.1007/s10257-022-00553-8; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] Confidence: high. The RPA literature shows that humans had been serving as the glue across disconnected systems, that automation moves more of that long-tail work into software agents, and that reliable scaling then depends on explicit exception handling, monitoring, privileged-access control, and change management.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4] Confidence: medium. Perrow's normal accident theory provides the clearest first-principles explanation for the gap because tighter coupling and interactive complexity make consequence propagation faster and harder to interrupt once human slack is removed.
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.iso.org/standard/65694.html] Confidence: medium. A framework provision that closed the gap would need to require organizations to inventory human-derived friction before autonomy increases and to prove that rate limits, approval thresholds, segmentation, monitoring, exception routing, and shutoff mechanisms replace the lost mitigation.

Evidence Map

Claim	Source	Confidence	Notes
[inference] No reviewed framework explicitly names human speed, attention, fatigue, or working hours as controls whose removal must trigger engineered replacement.	https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html	high	Absence claim limited to reviewed accessible texts and public summaries.
[inference] DORA is the closest adjacent framework because it requires continuous monitoring, automatic alerting, automated isolation, controlled change management, and interdependency-aware continuity planning without explicitly naming removed human rate limits.	https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554	medium	Strongest control-substitution analogue in reviewed framework set.
[inference] Basel Committee principles and APRA CPS 230 require explicit control environments, monitoring, remediation, and ICT risk management, but not the narrower implicit-controls mechanism.	https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230	high	Strong prudential support for critique, weak explicit naming of mechanism.
[inference] NIST AI RMF 1.0 requires designed human-AI oversight, monitoring, and safe decommissioning, but does not explicitly require identification of removed human operating constraints.	https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-ai-rmf-10	medium	Shows explicit oversight design without the narrower concept.
[fact] The Flash Crash record documents automated execution compressed to 20 minutes, liquidity-provider pauses, and post-event circuit breakers that recreated assessment time explicitly.	https://www.sec.gov/files/marketevents-report.pdf	medium	Primary analogue for speed-of-consequence and engineered brakes.
[inference] RPA evidence shows humans were previously acting as the glue across systems and that scaled automation requires explicit exception handling, monitoring, privileged-access control, and change management.	https://doi.org/10.1007/s12599-018-0542-4; https://link.springer.com/article/10.1007/s10257-022-00553-8; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists	high	Combines academic and practitioner evidence.
[inference] Perrow's normal accident theory best explains why removal of human slack increases blast radius in tightly coupled systems.	https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://www.sec.gov/files/marketevents-report.pdf	medium	Conceptual support rather than a framework citation.
[inference] A gap-closing framework clause would need to require inventory and replacement of human-derived friction before autonomy scales.	https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://www.iso.org/standard/65694.html	medium	Derived from reviewed duties plus analogue evidence.

Assumptions

[assumption; source: https://doi.org/10.1007/s12599-018-0542-4; https://doi.org/10.1007/s10257-020-00472-6] Assumption: Human speed, attention, fatigue, and working hours historically bounded citizen-development blast radius even though the reviewed corpus did not contain a direct pre-agentic measurement study. Justification: The claim is inferred from shadow-information-technology and automation analogues rather than from a direct quantitative historical study of citizen development specifically.
[assumption; source: https://academic.oup.com/rfs/article/32/5/2024/5428080; https://www.gartner.com/en/documents/3983454; https://www.iso.org/standard/65694.html; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final] Assumption: The explicit-gap conclusion is limited to the reviewed accessible framework texts and public summaries. Justification: Several seeded or granular sources were paywalled or inaccessible in this runtime, although no reviewed evidence suggested that those missing texts explicitly close the gap.

Analysis

[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://handbook.apra.gov.au/standard/cps-230; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final; https://www.iso.org/standard/65694.html] I weighted primary framework texts and official summaries more heavily than vendor or practitioner commentary when determining whether explicit coverage exists, so the explicit-gap conclusion is driven by regulator and standards documents rather than by later commentary.
[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4; https://www.isaca.org/resources/isaca-journal/issues/2023/volume-2/rpa-is-evolving-but-risk-still-exists] I used analogue evidence from high-frequency trading and RPA not to prove that the frameworks already contain the mechanism, but to test whether consequence speed, exception routing, and control substitution behave in practice the way the implicit-controls hypothesis predicts.
[inference; source: https://press.princeton.edu/books/paperback/9780691004129/normal-accidents; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554] Perrow's coupling model and DORA's explicit interdependency and continuity language align strongly enough that the first-principles argument is more credible than a pure novelty claim detached from operational-risk theory.
[inference; source: https://doi.org/10.1007/s10257-020-00472-6; https://doi.org/10.1007/s12599-018-0542-4] The most important trade-off in the evidence is between directness and relevance: the shadow-information-technology and RPA sources are highly relevant to workaround estates but indirect on the exact phrase "implicit controls," while the framework texts are direct on explicit controls but silent on the narrower mechanism.

Risks, Gaps, and Uncertainties

[assumption; source: https://doi.org/10.1007/s12599-018-0542-4; https://doi.org/10.1007/s10257-020-00472-6] Direct empirical studies that quantify human pace, attention, fatigue, or working hours as a measured blast-radius cap for citizen development were not found in the reviewed corpus.
[assumption; source: https://academic.oup.com/rfs/article/32/5/2024/5428080; https://www.gartner.com/en/documents/3983454] The inaccessible Oxford and Gartner sources may contain additional supporting detail for the technology-analogue section, but they do not change the reviewed-framework conclusion because the explicit-gap claim is grounded in accessible primary framework texts.
[assumption; source: https://www.iso.org/standard/65694.html; https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final] ISO 31000 and NIST SP 800-53 were only accessible at public-summary level in this runtime, so claims about their silence on the mechanism are lower precision than the DORA, Basel, APRA, NIST AI RMF, SEC, and RPA claims.

Open Questions

[inference; source: https://www.sec.gov/files/marketevents-report.pdf; https://doi.org/10.1007/s12599-018-0542-4] Which concrete control patterns best replace lost human friction in enterprise agent deployments, for example rate limits, mandatory approval thresholds, segmentation, exception-routing thresholds, or time-bounded credentials?
[inference; source: https://www.bis.org/fsi/fsisummaries/psmor.htm; https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022R2554; https://airc.nist.gov/airmf-resources/airmf/5-sec-core/] How should boards and risk committees quantify blast-radius change when a workflow moves from human execution to continuous agent execution under otherwise unchanged permissions and process maps?

Output

(Fill in when completing - what was produced as a result of this research?)

Type: knowledge
Description: Research note concluding that reviewed frameworks govern explicit controls around automation but do not explicitly name removal of implicit human rate-limiting as a control-substitution mechanism, with Flash Crash, RPA, and normal accident theory used to characterize the gap.
Links: