Public Comments

Status: Original and derived artifacts are available for this document.

Download: pdf

Artificial Intelligence for Corporate Governance in
the U.S. Defense Industrial Base: A
Scholar-Practitioner Framework for Accountability,
Ethics, and National Security
Satyadhar Joshi
Alumnus, International MBA, Bar-Ilan University, Israel
Alumnus, M.S. in Information Technology, Touro College, New York, USA
ORCID: 0009-0002-6011-5080
Email: [email protected]

Abstract—The U.S. defense industrial base confronts a confluence of governance challenges unprecedented in scope and
consequence: financial fraud, systemic product quality failures,
safety-critical software deficiencies, and supply chain opacity that
collectively jeopardize national security. Watershed events—the
Boeing 737 MAX disasters, persistent subcontracting irregularities, and recurrent Nunn-McCurdy cost breaches—illuminate
the structural inadequacy of traditional governance mechanisms
for an industry in which technical complexity, classified information flows, and sovereign risk intersect. This paper develops
a comprehensive, theory-grounded framework for integrating
Artificial Intelligence (AI) into the corporate governance architectures of U.S. defense contractors. Drawing on six canonical
governance theories—agency, stewardship, resource dependency,
stakeholder, institutional, and transaction cost economics—and
a systematic review of the emerging AI-governance literature,
we demonstrate how AI can operationalize ethical principles,
substantially reduce information asymmetry, enable continuous
auditing, and anticipate supply chain disruptions at scale. The
proposed AI governance meta-layer is calibrated to the specific
regulatory obligations imposed by the International Traffic in
Arms Regulations (ITAR), the Defense Federal Acquisition Regulation Supplement (DFARS), and the Paperwork Reduction Act
of 1995. Concurrently, this paper constitutes a formal public
comment submitted in response to Federal Register Volume 91,
Issue 103 (May 29, 2026), concerning the proposed extension of
information collection requirement OMB Control Number 07040253 under DFARS Part 244. The analysis demonstrates how
AI-enabled automation can enhance the quality, practical utility,
and analytical rigor of contractor purchasing system data while
materially reducing respondent burden—directly addressing the
four areas of public comment invited by the Defense Acquisition
Regulations System (DARS).
Index Terms—Artificial Intelligence, Corporate Governance,
Defense Industrial Base, National Security, Accountability,
Agency Theory, Stakeholder Theory, Ethical AI, Risk Management, Sarbanes-Oxley, DFARS, ITAR, Subcontracting Oversight,
Paperwork Reduction Act, Federal Register.

I. I NTRODUCTION : G OVERNANCE FAILURE AS A
NATIONAL S ECURITY R ISK
Corporate governance failures in the defense sector are
not merely matters of financial or reputational concern—
they constitute direct threats to national security. When a

missile guidance component is defective, when a satellite’s
attitude control software harbors an undetected edge-case fault,
or when cost overruns on a major acquisition program are
systematically concealed from congressional oversight, the
consequences extend far beyond the balance sheet. The U.S.
defense industrial base is unique among commercial sectors
in that its governance failures carry sovereign-level risk: they
can erode military readiness, embolden adversaries, and cost
lives.
Against this backdrop, the emergence of agentic AI in
2025–2026—autonomous systems capable of reasoning, planning, and adapting across complex operational environments—
presents a transformative opportunity for defense governance.
AI offers the prospect of a meta-governance layer: a persistent, real-time oversight infrastructure that monitors financial
integrity, manufacturing quality, regulatory compliance, and
supply chain resilience simultaneously, surfacing risk signals
that would otherwise remain invisible until a catastrophic
failure event.
This paper advances three interrelated contributions. First,
it synthesizes classical governance theory with the AIgovernance literature to produce a unified framework applicable across the defense industrial base—from prime contractors to Tier-3 subcontractors. Second, it draws generalizable
lessons from major corporate failures in defense-adjacent
industries, demonstrating how AI governance mechanisms
could have prevented or substantially mitigated each. Third,
it applies this framework to the specific context of DFARS
Part 244 contractor purchasing system oversight, with the dual
purpose of advancing scholarship and serving as a formal
public comment to DARS on the proposed extension of OMB
Control Number 0704-0253.
The central thesis is as follows: classical governance frameworks remain necessary but are no longer sufficient for the
complexity, pace, and stakes of modern defense contracting.
The strategic integration of AI does not displace human
judgment or board authority—it augments them, providing
a data-verified, continuously updated foundation upon which

sound governance decisions can be made. When embedded
within robust institutional structures and guided by explicit
ethical frameworks, AI can restore the defense industrial base’s
standing as a reliable steward of public resources and national
security.
A. Structure of This Paper
Section II situates the analysis within the specific context of
DFARS Part 244 and the Federal Register notice. Section III
reviews the foundational principles of defense contractor
governance. Section IV maps classical governance theories
to the defense context. Section V extracts AI governance
lessons from major corporate failures. Section VI maps AI
capabilities to each governance theory. Section VII addresses
the ethical imperative of preventing safety-critical failures.
Section VIII provides targeted recommendations on AI-driven
information collection improvement. Sections IX and X survey
benefits and barriers. Section XI offers board-level strategic
guidance. Section XII presents structured recommendations for
the Federal Register response, and Section XIII draws broader
policy implications.
B. Formal Response to Federal Register Notice DARS-20260202
This paper is submitted as a substantive public comment
in response to the Federal Register notice published May
29, 2026 (Volume 91, Issue 103), in which the Defense
Acquisition Regulations System announced the proposed extension of an information collection requirement under the
Paperwork Reduction Act of 1995, OMB Control Number
0704-0253, covering DFARS Subcontracting Policies and Procedures (DFARS Part 244). The current OMB approval expires
August 31, 2026; DoD proposes a three-year extension.
The notice invites public comment on four specific areas:
1) Whether the proposed collection of information is necessary for the proper performance of DoD functions,
including whether the information will have practical
utility;
2) The accuracy of DoD’s estimate of the burden of the
proposed information collection;
3) Ways to enhance the quality, utility, and clarity of the
information to be collected;
4) Ways to minimize the burden of the information collection on respondents, including through the use of
automated collection techniques or other forms of information technology.
This paper addresses all four areas. Points (iii) and (iv)
receive the most extensive treatment, as the authors believe
AI-driven automation represents the most consequential nearterm opportunity for both the DoD and the contractor community. Point (i) is addressed in Section II; point (ii) in
Section XII. The framework developed herein is, however,
deliberately broader than this single notice: it is intended as a
durable reference for any defense contractor, policymaker, or
governance practitioner seeking to integrate AI into defense
acquisition compliance.

II. T HE DFARS PART 244 I NFORMATION C OLLECTION
R EQUIREMENT: C ONTEXT AND I MPLICATIONS
A. Regulatory Architecture of Contractor Purchasing System
Oversight
DFARS Part 244 establishes the regulatory framework governing contractor purchasing system reviews (CPSRs). Two
provisions are central to the information collection requirement under review:
DFARS 244.305—Granting, Withholding, or Withdrawing Approval: Authorizes administrative contracting officers (ACOs) to evaluate the acceptability of a
contractor’s purchasing system upon completion of the
in-plant portion of a CPSR, to approve or disapprove
the system, and to pursue correction of identified material weaknesses. Disapproval triggers a requirement for
Government consent to individual subcontracts and may
precipitate financial withholding or the exercise of other
contractual remedies.
• DFARS Clause 252.244-7001—Contractor Purchasing
System Administration: Mandates that contractors respond in writing within 30 days of an ACO’s initial determination identifying material weaknesses. Following the
ACO’s evaluation of that response, if material weaknesses
persist, the contractor has 45 days to either remediate
the weaknesses or submit an acceptable corrective action
plan.
•

The current information collection covers 22 respondents,
generating 44 annual responses at an estimated average burden
of 8 hours per response, for a total annual burden of 372 hours
(OMB Control Number 0704-0253, as reported in the Federal
Register notice).
B. The Practical Utility of Purchasing System Oversight
The information collected under DFARS Part 244 is unambiguously necessary for the proper performance of DoD
oversight functions. Contractor purchasing systems govern the
acquisition of subcontracted goods and services that, in aggregate, constitute a substantial fraction of major defense program
costs. Deficiencies in these systems—inadequate competition among subcontractors, insufficient pricing analysis, noncompliance with flow-down requirements—translate directly
into cost overruns, schedule delays, and quality degradation
in delivered systems. The ACO’s ability to approve, condition,
or disapprove a contractor’s purchasing system is therefore a
foundational instrument of defense acquisition stewardship.
However, the practical utility of collected information is
constrained by the limitations of the current manual, episodic
review model. Information gathered at the time of a formal
CPSR may be stale by the time corrective actions are assessed.
Responses are of variable completeness and analytical depth.
And the 30- and 45-day response windows, while appropriate
as legal deadlines, may be insufficient for contractors to conduct the root-cause analysis necessary to produce genuinely
informative corrective action plans. AI offers a pathway to

transform this episodic, compliance-driven data collection into
a continuous, analytically rich oversight infrastructure.
C. Current Burden and Its Potential Understatement
The DoD’s estimate of 372 annual burden hours for 22
respondents warrants scrutiny. The 8-hour average per response may accurately reflect straightforward cases involving
a limited number of discrete material weaknesses. However,
for prime contractors operating complex, multi-tier supply
chains—where a purchasing system review may implicate
hundreds of subcontractor relationships across multiple commodity areas—the actual burden of preparing a substantive,
evidence-based response within 30 days is likely to be substantially higher. The estimate also does not appear to capture
indirect burden: the time spent by procurement, finance, and
legal personnel coordinating data collection across enterprise
systems, or the opportunity cost of diverting senior staff from
mission-critical activities during a response cycle.
This concern is not merely academic. An underestimated
burden figure may cause DoD to underweight the value of
investments in automated collection tools that could reduce
that burden, and may cause OMB to apply an inappropriately
low threshold in evaluating the proportionality of the collection
requirement. Section XII returns to this point with specific
recommendations.
III. F OUNDATIONAL P RINCIPLES OF G OVERNANCE IN THE
D EFENSE S ECTOR
Effective corporate governance in the defense industry is
distinguished from general commercial governance by the
sovereign nature of the primary customer, the classified character of much operational information, and the life-safety
consequences of component or system failure. Seven principles
are foundational.
Transparency demands open disclosure of material
information—board decisions, financial performance, and,
critically, quality or safety issues affecting systems destined for
military use [7]. In the defense context, transparency extends
to supply chains: contractors must be able to disclose the
provenance of rare-earth materials, electronic components, and
embedded software to prevent counterfeiting, unauthorized
substitution, or supply chain infiltration by adversarial actors.
Accountability establishes unambiguous chains of
responsibility—from the production floor through the
executive suite to the board of directors—and extends upward
to oversight bodies including the Defense Contract Audit
Agency (DCAA), the Defense Contract Management Agency
(DCMA), and congressional oversight committees [8]. For
defense contractors, accountability also flows downward
through the subcontract hierarchy, as prime contractors bear
regulatory responsibility for the purchasing systems and
compliance postures of their major subcontractors.
Fairness governs both internal corporate relationships—
preventing self-dealing among directors and protecting minority shareholders—and external procurement relationships,
ensuring that small businesses and disadvantaged firms receive

equitable access to subcontracting opportunities consistent
with FAR Part 19 requirements.
Responsibility demands ethical behavior that transcends
legal minimums. In defense manufacturing, this principle
carries particular moral weight: a component failure in a
weapon system or a military aircraft can result in the loss of
military personnel, civilian casualties, or mission failure with
geopolitical consequences [1]. Responsibility encompasses
both positive duties (producing reliable, fully tested systems)
and negative duties (refusing to conceal anomalies to preserve
contract performance incentives).
Risk management requires the systematic identification,
assessment, and mitigation of operational, financial, and compliance risks across a portfolio that may include exportcontrolled technologies, classified programs, sole-source supply relationships, and long-term fixed-price contracts subject
to Nunn-McCurdy breach reporting. Defense contractors face
a distinctive risk topology that general enterprise risk frameworks do not adequately address.
Compliance with ITAR, the Export Administration Regulations (EAR), DFARS, and applicable military standards (MILSTD) forms a non-negotiable governance foundation, enforced
through mechanisms including PCAOB oversight of publicly
traded contractors and DCAA audit authority [9].
Ethical culture—the shared values, norms, and behavioral expectations that shape decision-making throughout the
organization—must be actively and continuously cultivated
by senior leadership. The ”spiral of silence” that enabled
the Enron fraud and the Boeing safety culture failures operated through the same organizational dynamic: individuals
who detected problems self-censored because they perceived
organizational norms as hostile to dissent [2]. Board-level
governance mechanisms must actively disrupt this dynamic
through robust whistleblower protections, visible leadership
modeling, and genuine psychological safety for quality and
safety reporting.
IV. C LASSICAL G OVERNANCE T HEORIES AND T HEIR
A PPLICATION TO D EFENSE C ONTRACTING
Six classical governance theories provide the theoretical
infrastructure for the AI governance framework developed in
this paper. Each captures a distinct dimension of the principalagent dynamics, resource dependencies, and institutional pressures that characterize the defense contracting environment.
Agency theory identifies the principal-agent problem as the
central challenge of corporate governance: managers (agents)
may pursue self-interest at the expense of shareholders (principals) when information asymmetry and misaligned incentives
permit [11]. In defense firms, agency risks manifest acutely
in cost-plus contracting environments where the incentive
to control costs is weak, in the concealment of production
defects to avoid bonus clawbacks or contract terminations,
and in accounting manipulations that obscure true program
cost trajectories. The classified nature of many defense programs compounds information asymmetry, limiting the board’s
ability to independently verify management representations.

Stewardship theory offers a complementary perspective:
executives may be intrinsically motivated to act as stewards
of the organization’s long-term interests, seeking reputational
and mission-oriented rewards rather than short-term personal
gain [12]. The mission-driven character of national security
work creates favorable conditions for stewardship culture in
defense firms—provided that governance structures support,
rather than undermine, long-term orientation.
Resource dependency theory conceptualizes the board as a
boundary-spanning mechanism that provides access to critical
external resources: regulatory approvals, military contracts,
political relationships, and strategic supply chain partnerships
[12]. For defense contractors, the loss of a single major
program or a critical single-source supplier can pose existential
risks that dwarf those faced by diversified commercial firms.
Stakeholder theory extends the governance obligation beyond shareholders to encompass employees, local communities, taxpayers who fund the defense budget, and the military
personnel whose operational effectiveness and safety depend
on the quality of delivered systems [13]. The Department
of Defense itself must be understood as a stakeholder with
governance interests that differ in character from those of
commercial customers.
Institutional theory explains organizational behavior as
a response to coercive, normative, and mimetic pressures
from the external environment [14], [15]. Defense contractors operate within a dense institutional environment—
military standards, ITAR, DFARS, FAR, congressional oversight, and informal professional norms of the defense acquisition community—and must maintain legitimacy across all of
these dimensions simultaneously.
Transaction cost economics (TCE) analyzes governance
choices—including make-or-buy decisions and monitoring
architectures—as functions of asset specificity, uncertainty,
and transaction frequency [11]. Defense subcontracting relationships frequently involve highly asset-specific investments
(specialized tooling, classified facilities, unique technical expertise) that create lock-in and elevate the cost of switching
suppliers, with direct implications for purchasing system design and oversight.
V. L ESSONS FROM C ORPORATE G OVERNANCE FAILURES :
I MPLICATIONS FOR D EFENSE
The defense industry is not immune to the classes of governance failure that have produced catastrophic outcomes in
other sectors. An analysis of three canonical failures—Enron
(financial fraud), Kia/Hyundai (systematic product quality
concealment), and Boeing 737 MAX (safety-critical software
and cultural failure)—reveals structural vulnerabilities that
are directly present in defense contracting environments and
that AI governance mechanisms are specifically designed to
address.
A. Enron (2001): Financial Fraud and the Limits of Audit
The Enron collapse represents the archetype of financial
governance failure driven by information manipulation at

scale. Executives systematically deployed mark-to-market accounting, special purpose entity (SPE) structures, and deliberate disclosure obfuscation to conceal billions in debt
while inflating reported earnings [2]. The failure of Arthur
Andersen’s audit function, the passivity of the board’s audit
committee, and the complicity of external legal and financial
advisors demonstrated that periodic, human-conducted audits
are vulnerable to sophisticated, sustained manipulation. The
Sarbanes-Oxley Act of 2002 addressed some of these vulnerabilities through enhanced internal control requirements and
auditor independence rules, but the fundamental information
asymmetry between managers and overseers remains [8], [9].
For defense contractors, analogous risks manifest in longduration cost-plus programs where percentage-of-completion
accounting can be manipulated, in the use of intracompany
transfer pricing to shift costs between government and commercial programs, and in the underreporting of earned value
management (EVM) variances to avoid triggering contractual
remedies or congressional notification requirements. An AIdriven continuous audit function—analyzing journal entries,
related-party transactions, EVM data streams, and revenue
recognition patterns in real time—can detect anomalies that
periodic audits miss. Natural language processing applied to
internal communications can identify patterns of deliberate
obfuscation or ”tone at the top” that correlates with elevated
fraud risk, drawing on the established research utility of the
Enron email corpus as a training dataset [2].

B. Kia/Hyundai (2010–2020): Systematic Quality Concealment
The decade-long saga of Theta II engine failures in Kia
and Hyundai vehicles—involving engine seizure, uncontrolled
fires, and documented pressure to delay defect reporting to
regulators—illustrates the governance failure mode of systematic quality concealment under production schedule and
financial performance pressure. The parallel for defense manufacturers is direct: propulsion system anomalies in missiles
or satellite thrusters, structural defects in aircraft components,
or electromagnetic compatibility failures in electronic warfare
systems may be discovered during testing but suppressed to
maintain contract award schedules or avoid cost-plus adjustments.
AI-powered non-destructive testing (NDT), integrating computer vision analysis of X-ray, ultrasonic, and thermographic
imaging data, can detect microscopic material defects—
hairline fractures, inclusions, delaminations—that human inspectors miss at production rates [16]. Critically, AI systems
can maintain cryptographically integrity-protected logs of all
inspection results, preventing the retrospective deletion or
alteration of test records that characterized the Kia/Hyundai
concealment. Predictive maintenance algorithms applied to
production test stand telemetry can identify batch-level failure propensity before components are accepted into finished
assemblies.

C. Boeing 737 MAX (2018–2019): Safety-Critical Software
and Cultural Failure
The twin crashes of Lion Air Flight 610 and Ethiopian
Airlines Flight 302, killing 346 people, resulted from a
chain of governance failures that began with a software design decision—the Maneuvering Characteristics Augmentation
System’s reliance on a single angle-of-attack sensor—and
propagated through inadequate safety review, concealment of
MCAS functionality from pilots and the FAA, and a corporate
culture in which schedule pressure systematically overrode engineering rigor [1]. The Boeing case is particularly salient for
defense governance because the failure mode—autonomous
software behavior that was not fully understood by operators,
combined with management pressure that suppressed safety
dissent—is directly replicated in the development of autonomous weapon systems, AI-enabled flight control systems,
and cyber-physical defense platforms.
AI-augmented software verification, incorporating formal
methods, model checking, and systematic fuzz testing, can
enumerate edge cases in safety-critical code that human test
teams fail to anticipate. For a defense guidance system, AI
can simulate millions of sensor failure scenarios to verify that
the system degrades gracefully rather than catastrophically.
NLP applied to design review documents, change control
board minutes, and engineering email communications can
identify patterns of ”production pressure” language—schedule
commitments displacing test completion requirements—and
flag these as cultural leading indicators of safety risk. AIenforced change control workflows can ensure that no software
update to a fielded system is deployed without completing all
required safety reviews.
D. Cross-Cutting Synthesis: The Multi-Dimensional Governance Imperative
Defense contractors simultaneously face all three failure
modes: financial fraud risks (Enron-type) in complex contract accounting and supply chain pricing; product quality
risks (Kia-type) in precision-manufactured components where
defect rates must approach zero; and safety-critical software
risks (Boeing-type) in the autonomous and semi-autonomous
systems that increasingly define modern defense capabilities.
An effective AI governance architecture must address all three
dimensions within a unified oversight platform, enabling the
board to maintain a holistic risk picture rather than managing
each failure mode through separate, siloed processes.
VI. M APPING AI C APABILITIES TO C LASSICAL
G OVERNANCE T HEORIES
Each classical governance theory identified in Section IV
maps to specific AI capabilities that operationalize its prescriptions within the defense context. This mapping constitutes the
theoretical core of the proposed framework.
A. Agency Theory: AI as an Information Asymmetry Equalizer
Agency theory prescribes mechanisms that reduce information asymmetry between principals and agents. AI operationalizes this prescription through real-time, auditable data

streams that remove managerial control over what information
reaches the board. Machine learning algorithms continuously
monitor production metrics, financial transactions, purchasing
system compliance indicators, and executive decisions, feeding
verified data to board dashboards that bypass the management
reporting layer. For purchasing system oversight specifically,
AI systems analyzing DFARS Part 244 compliance metrics
can detect unauthorized subcontract placements, inadequate
competition documentation, or missing pricing analyses before they accumulate into reportable material weaknesses—
transforming the ACO’s oversight function from a periodic
audit into a continuous assurance regime [17].
B. Stewardship Theory: AI as a Long-Term Performance Verifier
Stewardship theory is operationalized through governance
mechanisms that align executive incentives with long-term
organizational health. AI supports stewardship by maintaining objective, tamper-resistant performance records that enable genuine long-term accountability. AI-enabled balanced
scorecards can track program performance against multi-year
delivery, quality, and cost objectives—providing the board
with verified data to evaluate whether executives are acting
as stewards of the firm’s defense mission or optimizing for
short-term compensation metrics [5]. NLP analysis of internal
communications can detect when managers are acting in genuine stewardship (proactively surfacing quality issues) versus
in self-interested concealment.
C. Resource Dependency Theory: AI as Strategic Intelligence
Resource dependency theory prescribes board-level attention to the external resource environment. AI transforms this
from a periodic board agenda item into a continuous intelligence function. Machine learning models analyzing geopolitical news, export control regulatory changes, supplier financial
health indicators, and commodity price trends provide early
warning of resource disruptions months before they materialize
in delivery failures or cost spikes [18]. For DFARS Part
244 compliance specifically, AI monitoring of subcontractor
financial health and performance metrics can identify suppliers
approaching insolvency or quality deterioration before they
trigger purchasing system material weaknesses.
D. Stakeholder Theory: AI as Continuous Compliance Monitor
Stakeholder theory prescribes governance mechanisms that
track and respond to the full range of stakeholder interests. AI
operationalizes this through continuous, multi-channel compliance and sentiment monitoring. NLP processes regulatory updates from DoD, DCAA, DCMA, and congressional sources;
analyzes public reporting and news media for emerging contractor reputation issues; and monitors internal whistleblower
channels for early signals of governance deterioration. For
defense contractors with significant environmental footprints
(propellant manufacturing, testing range operations), AI can
monitor regulatory filings and community sentiment simultaneously, enabling proactive stakeholder engagement [19].

E. Institutional Theory: AI as Regulatory Conformity Engine
Institutional theory prescribes mechanisms for maintaining legitimacy across the dense regulatory environment of
defense contracting. AI serves as a regulatory conformity
engine: trained on the complete corpus of applicable regulations, military standards, and agency guidance documents,
the system automatically identifies regulatory changes, maps
them to affected internal processes, and generates updated
compliance checklists. When new AI ethics guidelines for
autonomous weapon systems emerge, or when DFARS clauses
are amended, the governance AI immediately propagates the
required process updates throughout the organization, reducing
the lag between regulatory change and organizational response
that creates compliance gaps.
F. Transaction Cost Economics: AI as a Monitoring Cost
Reducer
TCE prescribes governance structures that minimize the
sum of transaction, monitoring, and enforcement costs. AI
dramatically reduces monitoring costs in multi-tier defense
supply chains by automating the continuous review of supplier quality data, delivery performance, financial health, and
regulatory compliance. For purchasing system oversight under
DFARS Part 244, AI can automatically verify that all required
pricing analyses, competition documentation, and flow-down
clause compliance records are present and complete for each
subcontract action—reducing the manual labor required for
CPSR preparation and ACO review simultaneously. Smart
contract mechanisms can further reduce enforcement costs
by automating payment upon verified delivery of conforming
items [20].
VII. P REVENTING THE K ILL I SSUE : AI AND THE E THICS
OF S AFETY-C RITICAL G OVERNANCE
The most severe governance failure in defense manufacturing is the one that directly costs lives: a defective missile
guidance component that causes a fratricide event, a satellite thruster failure that leaves a communications satellite
uncontrolled, or an autonomous system software fault that
results in civilian casualties. This ”kill issue”—the possibility
that a governance failure directly produces lethal outcomes—
imposes ethical obligations on defense contractor boards that
transcend ordinary commercial fiduciary duties.
A. AI-Driven Anomaly Detection for Safety-Critical Components
Modern machine learning algorithms, trained on historical
inspection and test data, can identify microscopic anomalies in manufacturing output that human inspectors miss at
production scale [16], [21]. Continuous analysis of X-ray
inspection images, vibration signatures, material composition
reports, and assembly process telemetry enables the automatic quarantine of suspect components before they enter
the finished goods inventory. Critically, all inspection events
are recorded in integrity-protected logs that cannot be altered
retrospectively—eliminating the possibility of the retrospective

concealment that characterized the Kia/Hyundai and Boeing
cases.
B. Predictive Risk Assessment for Catastrophic Failure Modes
Machine learning models trained on historical test data
and informed by physics-based failure mode analysis can
predict which production batches carry elevated in-field failure probability. For precision strike systems, even a subpercent undetected defect rate may carry unacceptable safety
consequences. AI-driven predictive analytics enables boards
to mandate reinspection or redesign cycles before fielding,
embodying the classical ethical principle of precautionary risk
management—the obligation to prevent foreseeable harm even
under residual uncertainty [4].
C. Operationalizing Classical Ethics Through AI Governance
Classical ethical frameworks, which have historically operated as aspirational guides rather than operational constraints,
become genuinely actionable through AI governance systems.
Kantian deontology, with its categorical imperative requiring
that moral rules be universalizable, is operationalized when AI
systems audit production decisions against a codified ruleset—
flagging any deviation from mandated safety protocols as a
deontological violation regardless of schedule or cost consequences [4], [6]. When a production supervisor overrides a
required safety test, the AI documents the deviation, alerts
the audit committee, and initiates an escalation workflow—
making rule violations visible and consequential in real time.
Utilitarian ethics is operationalized through AI’s capacity
to model the expected consequences of governance decisions
at scale. When a contractor faces the temptation to ship
a component with a marginal anomaly to avoid a contract
penalty, an AI governance system can quantify the probabilistic expected harm—integrating estimated failure probability,
operational context, and downstream consequences—against
the financial benefit of on-time delivery, enabling a genuinely
utility-informed board decision rather than one driven by shortterm financial pressure.
D. Explainability as an Ethical Governance Requirement
For defense governance applications, every AI-driven recommendation that affects safety or procurement must be explainable to non-specialist decision-makers. Techniques such
as SHAP (SHapley Additive exPlanations) and LIME (Local
Interpretable Model-Agnostic Explanations) enable AI systems to generate plain-language justifications for their outputs,
supporting the classical governance virtue of accountability:
the board can explain and defend its AI-informed decisions
to regulators, military customers, and the public [7]. Explainability is not merely a technical requirement—it is an
ethical one, ensuring that AI systems do not substitute opaque
algorithmic authority for the human judgment that governance
accountability requires.

VIII. AI S OLUTIONS FOR DFARS PART 244
I NFORMATION C OLLECTION
This section directly addresses the Federal Register notice’s
invitation for comment on ways to enhance the quality, utility,
and clarity of information collected under DFARS Part 244,
and ways to minimize respondent burden through automated
collection techniques.
A. Continuous Automated Monitoring as a Structural Alternative to Episodic Review
The most consequential AI contribution to DFARS Part
244 compliance is not incremental improvement to the existing episodic CPSR model but rather its structural transformation. An AI-driven continuous monitoring architecture
would deploy machine learning models—trained on CPSR
findings, DFARS regulatory text, and contractor purchasing
system data—to monitor purchasing system performance on
an ongoing basis. Key indicators such as competition rates
by commodity, pricing analysis completion rates, flow-down
clause compliance, and supplier qualification status would be
tracked in real time. Deviations from acceptable performance
thresholds would trigger automated alerts, enabling contractors
to initiate corrective action before anomalies accumulate into
the material weaknesses that trigger formal ACO determinations.
This approach addresses the Federal Register’s point (iii)—
enhancing quality and utility—by ensuring that the information available to ACOs is current and analytically rich, rather
than a point-in-time snapshot of a purchasing system’s state
at the moment of an infrequent review. It also addresses
point (iv)—minimizing burden—by enabling contractors to
identify and resolve issues continuously rather than mobilizing significant organizational resources to respond to formal
determinations under tight statutory deadlines.
B. AI-Assisted Response Generation for Formal Determinations
When an ACO issues an initial determination identifying
material weaknesses, the 30-day response window creates
significant organizational pressure, particularly for prime contractors with complex, multi-tier supply chains. AI can substantially reduce the time and effort required to produce a
compliant, substantive response by:
1) Automatically gathering and consolidating relevant evidence from enterprise resource planning (ERP), procurement, quality management, and contract management
systems;
2) Generating structured, clause-by-clause draft responses
that address each identified material weakness with
supporting data;
3) Identifying and surfacing corrective action precedents
from prior CPSR cycles;
4) Tracking response deadlines across multiple simultaneous determination cycles and alerting responsible officials to approaching cutoffs.

For final determinations requiring a 45-day corrective action
plan, AI can further assist by generating milestone-based remediation plans, assigning ownership of corrective actions across
organizational functions, and providing automated progress
tracking that ensures commitments are met within the regulatory window.
C. Enhancing Information Quality Through Standardization
and Validation
The Federal Register notice specifically invites comment
on ways to enhance the quality, utility, and clarity of collected
information. AI enables the following quality enhancements:
1) Machine-readable standardized reporting format: A
structured, machine-readable format for CPSR responses
would enable automated analysis and cross-contractor
benchmarking by DoD, substantially enhancing the analytical utility of the information collected. DoD should
consider developing a standard data schema for DFARS
Part 244 responses, analogous to the XBRL tagging
standards adopted for SEC financial filings.
2) Automated data validation: AI systems can crossreference reported information against source system
data—verifying that stated competition rates, pricing
analyses, and supplier qualification records are consistent with underlying transactional data—reducing the
risk of inadvertent error or deliberate misrepresentation.
3) Completeness verification: NLP models trained on the
regulatory text of DFARS Part 244 and clause 252.2447001 can verify that all required elements of a response
have been addressed before submission, reducing the
likelihood of incomplete responses that require additional ACO follow-up.
4) Predictive analytics for future weakness identification: AI analysis of historical CPSR findings across
the contractor community can identify patterns that
predict future material weaknesses, enabling proactive
DoD outreach to contractors whose purchasing system
indicators suggest emerging risk.
D. Minimizing Respondent Burden: Quantitative Analysis and
AI Solutions
The DoD’s estimate of 372 annual burden hours across 22
respondents (8 hours per response, 2 responses per respondent)
merits careful scrutiny. For contractors with straightforward,
single-element material weakness determinations, 8 hours may
be a reasonable estimate. However, for prime contractors managing thousands of annual subcontract actions across multiple
commodity areas and organizational divisions, the coordination effort required to gather evidence, draft a substantive
response, obtain legal review, and submit within 30 days likely
substantially exceeds this figure.
AI automation can reduce respondent burden through five
mechanisms:
1) Automated data aggregation: AI systems integrate data
from ERP, procurement, quality, and contract management systems, eliminating the manual data-gathering

phase that accounts for a significant portion of response
preparation time.
2) Intelligent document processing: NLP extracts relevant
compliance evidence from supplier documents, subcontract files, and pricing analyses, reducing the manual
review burden for response preparation.
3) Automated compliance checking: AI continuously verifies supplier certifications, pricing analysis completion,
and flow-down clause compliance against DFARS requirements, maintaining a perpetually current compliance inventory.
4) Self-service compliance portals: Web-based portals
enabling contractors to submit responses and track corrective action status electronically would reduce administrative overhead and enable DoD to process responses
more efficiently.
5) Proactive issue identification: By identifying emerging
weaknesses before they become material, AI continuous
monitoring reduces the frequency of formal determination cycles, directly reducing the total annual burden
across the contractor community.
A conservative estimate, based on comparable AI implementation outcomes in financial audit and regulatory compliance contexts, suggests that AI automation could reduce
respondent burden by 50–70% within three to five years of
implementation [16], [17]. DoD should consider conditioning
the three-year extension on a midterm evaluation of AIassisted compliance tool adoption rates and corresponding
burden reduction outcomes.
IX. O PPORTUNITIES AND B ENEFITS OF AI-D RIVEN
D EFENSE G OVERNANCE
The strategic case for AI-augmented governance in the
defense sector rests on seven categories of benefit that extend
well beyond DFARS compliance.
Enhanced transparency and real-time board oversight:
AI systems provide continuous monitoring of financial transactions, procurement cycles, and production quality, delivering
dynamic risk dashboards to boards of directors that replace
retrospective quarterly reports with forward-looking risk intelligence [20]. This shift from periodic to continuous oversight
is qualitatively transformative—it enables boards to intervene
before governance failures become material events rather than
after.
Improved regulatory compliance and audit efficiency:
Defense contractors navigating ITAR, EAR, DFARS, FAR,
and applicable MIL-STD simultaneously face a compliance
burden of exceptional complexity. AI automates compliance
checking against this regulatory matrix, reducing human error,
lowering internal audit costs, and enabling compliance professionals to focus on high-judgment activities rather than data
gathering [16], [21].
Early detection of fraud and misconduct: NLP applied to
internal communications, procurement decisions, and financial
transactions can identify patterns associated with bribery, false
certifications, kickback schemes, and unauthorized technology

transfers before they escalate into criminal liability or contract
termination events [2].
Strategic supply chain risk management: Machine learning models analyzing geopolitical dynamics, supplier financial
health, and commodity market conditions provide early warning of supply chain disruptions that, in defense programs with
long lead times and limited supplier alternatives, can cascade
into program delays with national security consequences [18].
Strengthened DoD customer relationships: Demonstrating AI-augmented governance and compliance capability to
the Department of Defense serves as a meaningful competitive
differentiator in source selections that increasingly evaluate
contractor management systems as part of past performance
and responsibility determinations.
Reduced cost of capital: Institutional investors increasingly incorporate governance quality into cost-of-equity assessments. A demonstrable commitment to AI-enhanced
accountability—particularly in a sector where governance failures carry reputational and legal consequences of exceptional
magnitude—can lower borrowing costs and attract capital from
ESG-oriented funds [22].
Operational efficiency gains: Automating manual audit,
compliance, and reporting tasks enables skilled technical and
professional personnel to be redirected from administrative
compliance activities to mission-critical engineering, program
management, and innovation functions [19].
X. C HALLENGES AND BARRIERS TO AI G OVERNANCE
I MPLEMENTATION
A rigorous assessment of AI governance opportunities
requires equal candor about implementation barriers. Five
categories of challenge merit specific attention in the defense
context.
Data quality and legacy system integration: Many established defense contractors operate manufacturing and financial systems that were not designed for AI integration.
Decades of component quality records, procurement data, and
financial ledgers may reside in incompatible formats across
disconnected legacy systems. The upfront investment required
to cleanse, integrate, and normalize this data is substantial
and must be realistically accounted for in any AI governance
business case [17].
Algorithmic explainability and procedural fairness: Defense procurement decisions frequently involve the exercise
of discretionary judgment by contracting officers and boards
of directors who bear legal and ethical accountability for
their decisions. Black-box AI recommendations that cannot be
explained in terms accessible to non-specialist decision-makers
are inconsistent with the procedural fairness requirements of
government contracting and the fiduciary duties of corporate
directors [7]. Governance AI deployments must prioritize
explainable AI (XAI) architectures from the design stage.
Cybersecurity and adversarial AI risk: An AI governance
system in a defense contractor environment is itself a highvalue intelligence target. A compromised AI system could
falsely certify defective components as compliant, generate

misleading audit trails, or exfiltrate sensitive program data.
Governance AI must be protected through zero-trust network
architectures, adversarial robustness testing, and continuous
anomaly monitoring of the AI system’s own outputs.
Regulatory uncertainty and liability allocation: No specific federal standards currently govern the use of AI in
defense contractor governance or DFARS compliance functions. The legal liability implications of erroneous AI audit
findings—particularly if a contractor relies on an AI system’s
assurance of purchasing system compliance and subsequently
receives a material weakness determination—are unresolved.
DoD and industry stakeholders should collaborate to develop
clear standards for AI system certification in governance applications and liability frameworks for AI-assisted compliance
decisions.
Workforce capability gaps and cultural resistance: Effective AI governance implementation requires data scientists, AI
ethicists, and procurement professionals capable of designing,
operating, and critically evaluating AI systems—a skill set that
most established defense contractors lack in substantial depth.
Moreover, long-tenured procurement and finance managers
may perceive AI oversight as a threat to their authority,
creating cultural resistance that must be managed through
deliberate change leadership [10]. Governance AI should be
framed to stakeholders as augmenting, rather than supplanting,
professional judgment and expertise.
XI. S TRATEGIC C ONSIDERATIONS FOR D EFENSE
C ONTRACTOR L EADERSHIP
Six strategic considerations should inform board-level deliberation on AI governance investment in the defense context.
Board composition and AI literacy: Effective oversight of
an AI-augmented governance system requires board members
who can critically evaluate AI system outputs, understand their
limitations, and make informed judgments about algorithmic
recommendations. Defense contractor boards should consider
the appointment of directors with technology, cybersecurity, or
AI governance expertise—a qualification that research identifies as increasingly important for maintaining competitive
governance legitimacy [14], [15].
Ethical AI framework and institutional commitment:
Defense contractors should adopt a principled AI ethics
framework—such as the circular model emphasizing transparency, accountability, fairness, and inclusivity—and institutionalize it through governance policies, training requirements,
and audit mechanisms [4]. A ”Digital Oath” for AI systems,
binding governance AI to explicit ”Do No Harm” principles
and requiring human review of all recommendations that
affect safety or legal compliance, can provide meaningful
institutional assurance to DoD customers and investors [6].
Whistleblower protection and the limits of AI monitoring: AI monitoring of procurement and financial systems
must be carefully demarcated from employee surveillance.
Governance AI that processes internal communications for
fraud detection must be designed so that legitimate safety and

compliance reporting through protected whistleblower channels is not chilled by the perception of pervasive monitoring.
Sarbanes-Oxley whistleblower protections are non-negotiable
constraints on governance AI system design [9].
Phased implementation with high-risk area prioritization: Rather than enterprise-wide simultaneous deployment,
defense contractors should implement AI governance tools
in a phased sequence, beginning with the areas of highest
risk and regulatory consequence—such as supplier invoice
verification, purchasing system compliance monitoring, and
non-destructive testing image analysis—before expanding to
lower-risk applications [19]. This approach manages implementation risk, enables organizational learning, and generates
early evidence of return on investment.
Liability, insurance, and contract implications: Boards
must ensure that governance AI deployment does not inadvertently create unmanaged legal exposure. If an AI system
fails to detect a defect that subsequently causes program
failure or personal injury, the allocation of liability between
the contractor, the AI system vendor, and the DoD customer
requires advance legal analysis. Directors’ and officers’ insurance policies should be reviewed to confirm coverage for
AI-assisted governance decisions.
Competitive positioning and first-mover considerations:
The defense contractor that establishes a credible, DoDvalidated AI governance capability may realize significant
competitive advantages in source selections that evaluate contractor management systems. However, first-mover costs in
uncharted regulatory territory are real, and leadership must
weigh these against the long-term strategic value of governance leadership [20].
XII. S TRUCTURED R ECOMMENDATIONS : F ORMAL
R ESPONSE TO OMB C ONTROL N UMBER 0704-0253
The following recommendations are submitted in direct
response to the four areas of public comment invited by the
Federal Register notice regarding OMB Control Number 07040253.
A. Area (i): Necessity and Practical Utility of the Information
Collection
The information collection under DFARS Part 244 is necessary for the proper performance of DoD oversight functions
and should be extended. Contractor purchasing systems govern
subcontracted expenditures that represent a substantial fraction
of major defense program costs, and ACO oversight of these
systems is a foundational instrument of defense acquisition
stewardship. However, the practical utility of the information
collected can be substantially enhanced through AI-enabled
continuous monitoring, standardized machine-readable reporting, and predictive analytics. The authors recommend that
DoD condition the three-year extension on a structured evaluation of AI-assisted compliance tool adoption and its impact
on information quality and practical utility.

B. Area (ii): Accuracy of Burden Estimates
The DoD’s estimate of 372 annual burden hours (22 respondents, 2 responses each, 8 hours per response) is plausible
for straightforward cases but is likely to understate the true
burden for prime contractors managing complex, multi-tier
supply chains. The authors recommend that DoD conduct
a supplemental burden analysis stratified by contractor size
and purchasing system complexity, and that this analysis be
used to refine burden estimates in the three-year extension
period. Simultaneously, DoD should establish baseline metrics
for current burden levels to enable rigorous evaluation of AIdriven burden reduction over the extension period.
C. Area (iii): Enhancing Quality, Utility, and Clarity of Information
The authors recommend the following specific enhancements:
1) Develop a standardized, machine-readable data schema
for DFARS Part 244 responses, modeled on XBRL financial reporting standards, to enable automated analysis
and cross-contractor performance benchmarking;
2) Establish DoD-approved AI validation tools that contractors can use to verify the completeness and consistency
of their responses prior to submission;
3) Create a centralized DoD analytics capability to extract
systemic purchasing system risk patterns from the information collected, enhancing its value as a defense
acquisition intelligence resource;
4) Implement a pilot program to evaluate AI-driven continuous monitoring as an adjunct to the existing episodic
CPSR model, selecting 5–10 volunteer prime contractors
across diverse defense industry segments to test and
validate the approach before broader implementation.
D. Area (iv): Minimizing Respondent Burden Through Automation
The authors recommend the following:
1) DoD should issue guidance endorsing the use of AIassisted tools for DFARS Part 244 compliance monitoring and response preparation, providing clarity on the
permissibility of AI-generated draft responses subject to
human review and certification;
2) DoD should develop and make available a reference
architecture for AI-enabled purchasing system monitoring, enabling smaller contractors without in-house AI
capability to adopt standardized solutions;
3) The OMB burden estimate for the three-year extension
period should incorporate a phased reduction trajectory
reflecting projected AI automation adoption, with target
burden reductions of 30% by year two and 50–60% by
year three;
4) DoD should establish electronic self-service portals for
CPSR response submission and corrective action tracking, replacing paper-based processes that add administrative overhead without enhancing information quality.

XIII. C ONCLUSION AND P OLICY I MPLICATIONS
This paper has developed a comprehensive, theory-grounded
framework for AI-enhanced corporate governance in the U.S.
defense industrial base, with direct application to the information collection requirements under DFARS Part 244.
The foundational principles of transparency, accountability,
fairness, responsibility, risk management, compliance, and
ethical culture remain the bedrock of sound defense contractor
governance. The contribution of this framework is to demonstrate, with theoretical rigor and practical specificity, how AI
can operationalize these principles at a scale and speed that
human governance processes alone cannot match.
The mapping of AI capabilities to classical governance
theories reveals a coherent logic: agency theory’s prescription
for reduced information asymmetry is fulfilled by AI’s continuous audit function; stewardship theory’s orientation toward
long-term value is supported by AI’s verified performance
metrics; resource dependency theory’s emphasis on strategic
external intelligence is operationalized by AI’s supply chain
monitoring capability; stakeholder theory’s multi-constituency
accountability is served by AI’s continuous compliance monitoring; institutional theory’s demand for regulatory legitimacy
is addressed by AI’s conformity engine; and transaction cost
economics’ efficiency prescription is met by AI’s dramatic
reduction in monitoring costs.
The lessons of Enron, Kia/Hyundai, and Boeing 737 MAX
are not merely historical cautionary tales. They are structural
failure patterns that are directly replicable in defense contracting environments—and that AI governance mechanisms are
specifically designed to detect and prevent. The ”kill issue” in
defense manufacturing demands governance mechanisms that
can achieve near-zero-defect assurance at production scale, and
AI-augmented inspection, testing, and quality management
systems represent the most credible available pathway to that
standard.
For policymakers, this analysis suggests five priority actions:
1) Develop specific AI governance standards for defense
contractors under DFARS, including certification requirements for AI systems used in safety-critical manufacturing and compliance functions;
2) Establish a DoD-sponsored pilot program for AI-driven
purchasing system monitoring under DFARS Part 244,
using the OMB Control Number 0704-0253 extension
period as the implementation window;
3) Fund research on explainable AI architectures specifically designed for defense audit and compliance applications, ensuring that AI governance tools are compatible
with the procedural fairness requirements of government
contracting;
4) Create protected information-sharing mechanisms for AI
governance best practices across the defense industrial
base, enabling smaller contractors to benefit from prime
contractor implementation experience;

5) Develop clear regulatory guidance on the liability implications of AI-assisted compliance decisions, providing
the legal clarity that contractor boards require to authorize AI governance investments.
For defense contractor boards and senior leadership, the path
forward requires: deliberate AI literacy development at the
board level; phased implementation beginning with highestrisk governance domains; robust cybersecurity protection of
AI governance systems; and explicit commitment to the ethical
AI frameworks that align AI governance tools with the broader
obligations of responsible defense industrial citizenship.
The central thesis stands: AI is not a governance panacea,
and it cannot substitute for the human judgment, institutional
integrity, and ethical commitment that effective defense contractor governance requires. It is, however, a powerful and increasingly necessary meta-governance layer. When embedded
within robust traditional governance structures, calibrated to
the specific regulatory and ethical obligations of the defense
sector, and guided by the classical governance theories that
have defined the discipline, AI can help defense contractors fulfill their role as reliable, accountable, and ethically
grounded partners in U.S. national security—while simultaneously reducing the administrative burden that regulatory
compliance imposes on the contractor community and the
Government alike.
F ORMAL C OMMENT S UBMISSION S TATEMENT
This paper is submitted as a formal public comment in
response to Federal Register Volume 91, Issue 103 (May 29,
2026), Federal Register Document No. 2026-10731, regarding
the proposed extension of information collection requirement
OMB Control Number 0704-0253, Defense Federal Acquisition Regulation Supplement, Subcontracting Policies and
Procedures, Docket Number DARS-2026-0202. The analysis
and recommendations herein address all four areas of public
comment invited by the Defense Acquisition Regulations System. Comments may also be submitted electronically at https:
//www.regulations.gov or by email to [email protected], referencing OMB Control Number 0704-0253,
by the July 28, 2026 comment deadline.
D ECLARATION
The views expressed are those of the author and do not
represent any affiliated institution. This work is conducted
as independent research. This paper is a conceptual and
review contribution; all proposals and findings are derived
from and grounded in the cited literature. The author does not
claim novel experimental results but advances an integrated
framework synthesizing existing research for application to
defense governance and regulatory policy.
Portions of this manuscript were drafted with the assistance
of AI writing tools to improve clarity and organization. All AIgenerated content was reviewed, edited, and verified by the
author for coherence and factual accuracy; however, readers
are encouraged to independently verify cited claims. The LATEX
code was developed with the assistance of AI tools.

R EFERENCES
[1] R. Desai, “The Boeing 737 MAX Crisis: Public Trust, Aviation
Safety, and Corporate Accountability?” pp. 463–469. [Online].
Available: https://ajosr.org/wp-content/uploads/journal/published paper/
volume-3/issue-6/ajsr2025 nFGEQxgg.pdf
[2] K. Taylor-Neu, A. S. Rahaman, G. D. Saxton, and D. Neu,
“Tone at the top, corporate irresponsibility and the Enron emails,”
vol. 37, no. 9, pp. 336–364. [Online]. Available: https://doi.org/10.
1108/AAAJ-12-2023-6792
[3] J.-E. Pozner, A. Mohliver, and C. Moore, “Shine a Light: How
Firm Responses to Announcing Earnings Restatements Changed After
Sarbanes-Oxley,” vol. 160, no. 2, pp. 427–443. [Online]. Available:
https://www.jstor.org/stable/45278228
[4] S. A. Hosseini Tabaghdehi and O. Ayaz, “AI ethics in action
a circular model for transparency, accountability and inclusivity.”
[Online]. Available: https://doi.org/10.1108/JMP-03-2024-0177
[5] P. K., G. Sehgal, U. M. Padwal, D. Shah, D. Tewari, and
S. Rana, “Corporate Governance and Ethical Management Practices:
Building Trust in Modern Enterprises.” vol. 2, no. 4, pp. 5388–5396.
[Online]. Available: https://research.ebsco.com/linkprocessor/plink?id=
935212d8-a2e3-35c6-9b0b-c492e152e483
[6] M. Najm and p. u. family=Amaral Giovani, given=Joao Francisco, “ICAIMT — The Digital Oath for Corporate
Governance: A First Model.” vol. 24, no. 6, pp. 1–17.
[Online]. Available: https://research.ebsco.com/linkprocessor/plink?id=
799ff2b9-d658-35fa-ae6c-8b389b2ffacb
[7] C. Li, “AI-Driven Governance: Enhancing Transparency and
Accountability in Public Administration,” vol. 1, no. 1, pp. 1–
16. [Online]. Available: https://www.dsvg.cc/dsvg/article/view/1
[8] M. W. PEREGRINE, “The Continuing Governance Relevance of
Sarbanes-Oxley: Two decades on, Sarbanes-Oxley remains a touchstone
for ethics, oversight and the modern director’s duty.” vol. 50, no. 2, pp.
58–61. [Online]. Available: https://research.ebsco.com/linkprocessor/
plink?id=b75aab09-47c1-3509-b504-67e0a3b825ca
[9] D. L. King and C. J. Case, “SARBANES-OXLEY ACT
AND THE PUBLIC COMPANY ACCOUNTING OVERSIGHT
BOARD’S FIRST ELEVEN YEARS.” vol. 7, no. 1, pp. 11–22.
[Online]. Available: https://research.ebsco.com/linkprocessor/plink?id=
ef4fb984-b875-33a6-bc0b-630ad8e62abd
[10] I. Gallego-Álvarez and L. Rodriguez-Dominguez, “The Role of Organisational Factors and Corporate Governance in Business Ethics Practices.”
vol. 8, no. 1, pp. 1–17. [Online]. Available: https://research.ebsco.com/
linkprocessor/plink?id=8223f3c1-b335-39a4-b094-57b06d8d3adc
[11] C. D. Cole and R. Schneider, “Corporate Governance and
Capital Structure: Extending Agency Theory From Executives
to the Board of Directors.” vol. 20, no. 6, pp. 108–124.
[Online]. Available: https://research.ebsco.com/linkprocessor/plink?id=
617bf704-7fd1-317f-be8e-f88a67a878b3
[12] P. M. Madhani, “Diverse Roles of Corporate Board: A Review of
Various Corporate Governance Theories.” vol. 16, no. 2, pp. 7–28.
[Online]. Available: https://research.ebsco.com/linkprocessor/plink?id=
3e2f85c2-14db-36c5-9716-9a07d48c4fc8
[13] J. Amis, J. Barney, J. T. Mahoney, and H. Wang, “From the
Editors—Why We Need a Theory of Stakeholder Governance—And
Why This is a Hard Problem,” vol. 45, no. 3, pp. 499–503. [Online].
Available: https://journals.aom.org/doi/10.5465/amr.2020.0181
[14] M. A. Witt, S. Fainshmidt, and R. V. Aguilera, “Our Board,
Our Rules: Nonconformity to Global Corporate Governance Norms,”
vol. 67, no. 1, pp. 131–166. [Online]. Available: https://doi.org/10.
1177/00018392211022726
[15] ——, “Our Board, Our Rules: Nonconformity to Global
Corporate Governance Norms.” vol. 67, no. 1, pp. 131–166.
[Online]. Available: https://research.ebsco.com/linkprocessor/plink?id=
dd5c7ebc-73e3-3217-a05e-a5672f512b0a
[16] Olubusola Odeyemi, Kehinde Feranmi Awonuga, Noluthando
Zamanjomane Mhlongo, Ndubuisi Leonard Ndubuisi, Funmilola
Olatundun Olatoye, and Andrew Ifesinachi Daraojimba, “The role of AI
in transforming auditing practices: A global perspective review,” vol. 21,
no. 2, pp. 359–370. [Online]. Available: https://wjarr.com/node/4438
[17] H. H. Samara, H. A. Qudah, H. J. Mohsin, S. Abualhijad, L. Y.
Bani Hani, S. Al Rahamneh, and M. Z. AlQudah, “Artificial
intelligence and machine learning in corporate governance: A

bibliometric analysis,” vol. 44, no. 2, pp. 349–375. [Online]. Available:
https://journals.sagepub.com/action/showAbstract
[18] M. Becht, J. Mejino-López, and G. B. Wolff, “Who Controls the Defence
Industry?” [Online]. Available: https://www.jstor.org/stable/resrep79335
[19] A. K. Ulfah, K. M. Shavkatovna, M. N. Nuriddinovna, O. L.
Erkinovna, R. M. A. O’g’li, and A. Hidayati, “Transformation of
good corporate governance through AI: Innovations in enhancing
supervision and accountability.” vol. 3393, no. 1, pp. 1–21.
[Online]. Available: https://research.ebsco.com/linkprocessor/plink?id=
ea17a16a-7cf1-35a4-a95e-04fa1b1e1bf5
[20] O. S. Shaban and A. Omoush, “AI-Driven Financial Transparency
and Corporate Governance: Enhancing Accounting Practices with
Evidence from Jordan,” vol. 17, no. 9. [Online]. Available: https:
//www.mdpi.com/2071-1050/17/9/3818
[21] F. A. Wassie and L. P. Lakatos, “Artificial intelligence and the future of
the internal audit function,” vol. 11, no. 1, p. 386. [Online]. Available:
https://www.nature.com/articles/s41599-024-02905-w
[22] S. Bamel, Dipika, and S. Singh, “Mapping the research landscape
of business ethics, corporate governance, and corporate social
responsibility: A bibliometric and citation analysis.” vol. 29, no. 1, pp.
31–57. [Online]. Available: https://research.ebsco.com/linkprocessor/
plink?id=e83e58ab-f59c-3583-9409-2b5bb8c43f8d