NIST's Cyber AI Profile: What Boards Need to Know Before the August 2026 Deadline

NIST published the preliminary draft of IR 8596, the Cybersecurity Framework Profile for Artificial Intelligence (Cyber AI Profile), on December 16, 2025.

What follows is the full blog post.

---

Your board approved your organization's AI strategy. Your teams are deploying AI tools into production. Your customers are interacting with AI-powered systems. And somewhere between the excitement of competitive advantage and the reality of operational complexity, a critical question has gone unanswered: who owns the cybersecurity risk that AI just introduced?

A December 2025 McKinsey report underscores the governance gap: while more than 88% of organizations report using AI in at least one business function, only 39% of Fortune 100 companies disclosed any form of board oversight of AI.

That gap is not an oversight. It is a liability. Every week that passes without a structured AI cybersecurity governance program is a week your organization accumulates unquantified, uninsured, and undefended risk. NIST has handed boards a framework to close that gap. The question now is whether boards will act before the regulatory clock runs out.

A global survey of directors found that 66 percent report their boards have "limited to no knowledge or experience" with AI, and nearly one in three say AI does not even appear on their agendas.

That statistic should alarm every CEO forwarding this to their board chair. When the threat evolves faster than board fluency, the organization does not catch up gradually. It catches up through a breach.

The NIST Cyber AI Profile: What It Is and Why It Exists

On December 16, 2025, the National Institute of Standards and Technology (NIST) released its preliminary draft Cyber AI Profile (NIST IR 8596, Cybersecurity Framework Profile for Artificial Intelligence), a framework intended to provide organizations with guidance on managing AI-related risks.

This is not a bolt-on document or a standalone standard.

Rather than introducing an entirely new AI security framework, it leverages the existing Cybersecurity Framework (CSF) 2.0, signaling that AI risk is now inseparable from enterprise cyber risk management.

This document represents a year of collaboration with over 6,500 contributors from government, academia, and industry, and it maps AI-specific cybersecurity considerations onto the familiar CSF 2.0 structure that security teams already use.

For boards that have already invested in CSF alignment, this is not a new burden. It is an extension of existing governance architecture into AI territory.

Following the 45-day comment period, NIST plans to develop the initial public draft for release in 2026.

Final guidance is expected by the end of 2026, which means boards have a defined, finite window to prepare before the framework carries real regulatory weight. Waiting for the final version before acting is precisely the wrong strategy.

The Three Focus Areas Boards Must Understand

The Cyber AI Profile centers on three focus areas: securing AI systems (identifying cybersecurity challenges when integrating AI into organizational ecosystems and infrastructure); conducting AI-enabled cyber defense (identifying opportunities to use AI to enhance cybersecurity, and understanding challenges when using AI to support defensive operations); and thwarting AI-enabled cyberattacks (building resilience to protect against new AI-enabled threats).

Each focus area carries distinct board-level implications. Here is a framework-aligned view of what each demands from leadership:

When finalized, the profile will help organizations incorporate AI into their cybersecurity planning by suggesting key actions to prioritize, highlighting special considerations from specific parts of the CSF when considering AI, and providing mappings to other NIST resources, including the AI Risk Management Framework.

The Threat Data Boards Cannot Ignore

The adversary side of this equation is not speculative.

AI-generated phishing increased 1,265% since ChatGPT's launch in late 2022, and the average cost of a phishing-related breach now stands at $4.88 million, per IBM's 2025 data.

Deepfake incidents increased 680% year-over-year, with Q1 2025 alone recording 179 separate incidents.

78 percent of CISOs surveyed say that AI-powered threats are having a significant impact on their organizations, a 5 percent increase from 2024.

These are not technology statistics. They are financial exposure metrics. A board that cannot translate these numbers into D&O liability discussion items has a governance gap, not just a technology gap.

The rapid AI adoption has dramatically accelerated the speed, scale, and sophistication of cyber threats, far outpacing current enterprise cyber defenses, and more than three-quarters (77%) of organizations lack the essential data and AI security practices needed to protect critical business models, data pipelines, and cloud infrastructure.

For boards asking whether their organization is in the majority or minority, that 77% figure suggests the odds are not in their favor without deliberate action.

Boards that treat AI cybersecurity governance as a CISO problem rather than a fiduciary responsibility are misreading both the threat and the regulatory environment. When an AI system fails, investigators will ask what the board knew, when they knew it, and what governance structure they had in place. "I delegated that to IT" is not a defensible answer.

The Regulatory Convergence Boards Cannot Outrun

The Cyber AI Profile does not exist in a vacuum. It sits at the intersection of a converging regulatory environment that creates simultaneous obligations across multiple jurisdictions.

The EU AI Act entered into force on August 1, 2024, and will be fully applicable two years later on August 2, 2026, with prohibited AI practices and AI literacy obligations having entered into application from February 2025, and governance rules and obligations for GPAI models becoming applicable on August 2, 2025.

Many other obligations, including the comprehensive compliance framework for high-risk AI systems, are scheduled to apply from August 2, 2026.

The SEC's 2026 examination priorities have elevated cybersecurity and AI concerns above cryptocurrency, which dominated regulatory attention for the previous five years.

In the absence of defined, AI-specific disclosure guidance, public companies may increasingly rely on tools such as the Cyber AI Profile to contextualize AI-powered risk as part of broader cybersecurity risk management and governance programs, including in connection with Form 10-K and other public disclosures.

More than a third (36%) of companies now disclose AI as a separate 10-K risk factor, up from 14% last year.

Investors and regulators are reading those disclosures with increasing scrutiny. The Cyber AI Profile gives organizations a defensible framework to substantiate those disclosures with actual program evidence, not just boilerplate risk language.

Like the CSF, the Cyber AI Profile is voluntary for most organizations; however, organizations that align their risk management practices to these resources tend to be viewed by customers, investors, and regulators as more secure, resilient, and responsible.

Voluntary today does not mean inconsequential today. Framework alignment is becoming a competitive differentiator in enterprise sales cycles, M&A due diligence, and cyber insurance underwriting.

Framework Alignment: Where the Cyber AI Profile Fits

The Cyber AI Profile is structured around NIST CSF 2.0's six core functions: Govern, Identify, Protect, Detect, Respond, and Recover. For boards already familiar with these functions, the Profile adds AI-specific subcategories and priority ratings within each function, allowing organizations to triage their most urgent AI security gaps without rebuilding their entire governance architecture.

The Profile also intersects directly with complementary standards that boards should ensure their organizations are tracking:

NIST AI RMF: Addresses AI-specific risks including bias, explainability, and model robustness. The Cyber AI Profile and AI RMF are designed to be used together.

ISO 27001:2022: Information security management requirements that now include clauses directly relevant to AI supplier risk and access control for automated systems.

CIS Controls v8: Provides implementation guidance for the technical security controls that underpin both securing AI systems and defending against AI-enabled attacks.

CISA AI Cybersecurity Collaboration Playbook: Offers sector-specific guidance for organizations operating critical infrastructure with AI integration.

Nearly 3 in 4 (73%) companies now disclose alignment with an external framework such as NIST CSF 2.0, ISO 27001, or something similar, up from 57% last year and just 4% in 2019.

The market has voted. Framework alignment is now table stakes, not a differentiator. The question is whether your AI cybersecurity posture keeps pace with your AI deployment posture.

Emerging Trends Boards Must Monitor Before Final Guidance Arrives

Agentic AI Expands the Attack Surface Exponentially

The preliminary Cyber AI Profile was drafted during a period when most organizational AI consisted of predictive models and generative tools. Agentic AI systems, those that autonomously execute multi-step tasks, make decisions, and interact with external systems, represent the next frontier of both business value and security exposure.

AI has moved beyond predictive models into "agentic" systems that act on their own, execute tasks, and make decisions under human oversight. This evolution creates both opportunity and exposure, and boards are being asked to guide how organizations adopt these systems responsibly while still realizing business value.

The final Cyber AI Profile will need to address agentic systems more explicitly. Boards should ensure their security teams are already stress-testing governance controls against agentic AI deployments.

AI Is Becoming a Defense Asset, Not Just a Risk

The security ROI case for structured AI governance is increasingly clear.

88% of organizations that deployed agentic AI for security use cases reported a positive ROI, with gains including an 85% improvement in threat identification and a 65% reduction in time to resolution.

Boards that position the Cyber AI Profile adoption purely as a compliance exercise miss the operational performance dividend. AI-enabled defense, one of the Profile's three focus areas, is a measurable investment in reducing mean time to detect and respond.

Regulatory Convergence Will Create Audit Pressure in 2026

The confluence of AI and cybersecurity was among the priorities specifically noted in the SEC's recently published examination priorities and FINRA's Annual Regulatory Oversight Report.

Combined with the EU AI Act's August 2026 full applicability date, organizations with global operations face simultaneous audit pressure from multiple regulatory bodies.

Covered entities should ensure that they maintain AI risk governance processes that appropriately treat AI as one component of their bigger-picture, integrated cybersecurity risk framework.

The Cyber AI Profile provides the organizing structure to meet that expectation with evidence, not assertions.

COSAiS: The Control Companion Boards Should Know

Running parallel to the Cyber AI Profile is the SP 800-53 Control Overlays for Securing AI Systems (COSAiS).

NIST's NCCoE invited participants to hear updates regarding the SP 800-53 Control Overlays for Securing AI Systems (COSAiS) alongside discussion of the Cyber AI Profile.

COSAiS translates the Profile's strategic guidance into specific NIST SP 800-53 security controls, giving security teams the technical implementation roadmap that boards approve at a strategic level. Organizations using both documents together will be better positioned when regulators and auditors ask for evidence of AI cybersecurity controls.

Board Readiness Assessment: Where Does Your Organization Stand?

Use this checklist to assess your board's current AI cybersecurity governance maturity before the final Cyber AI Profile arrives:

Governance Foundation

[ ] The board has formally assigned AI cybersecurity oversight to a specific committee or designated director

[ ] AI systems are inventoried with associated cybersecurity risk classifications

[ ] AI vendor contracts include security requirements and right-to-audit clauses

[ ] The organization has an AI acceptable use policy approved at the executive level

Threat Posture

[ ] Security tabletop exercises include AI-enabled attack scenarios (deepfakes, AI-generated phishing, model poisoning)

[ ] The organization has tested its incident response plan against an AI-system-specific breach scenario

[ ] Third-party AI tools and APIs are included in the vendor risk management program

Regulatory Readiness

[ ] The organization has mapped its AI use cases against EU AI Act risk categories ahead of the August 2026 enforcement date

[ ] AI-related cybersecurity risks are disclosed in 10-K filings with framework-specific substantiation

[ ] The security team has reviewed the NIST IR 8596 preliminary draft and identified gaps against current controls

Strategic Alignment

[ ] The CISO has briefed the board on the Cyber AI Profile's three focus areas and their business implications

[ ] Budget has been allocated to address AI-specific security gaps identified against CSF 2.0 outcomes

[ ] The organization participates in NIST's NCCoE Cyber AI Community of Interest or equivalent sector-specific AI security working groups

A board scoring fewer than eight of these items has material governance gaps that regulators, insurers, and adversaries will eventually find.

How I Help

AI governance is where I focus the majority of my client work, and for good reason: it is the area where the gap between executive awareness and operational readiness is widest. I help organizations implement the NIST AI RMF alongside the Cyber AI Profile, assess OWASP LLM Top 10 risks across their deployed AI systems, build EU AI Act compliance roadmaps ahead of the August 2026 enforcement date, and translate those frameworks into board-level reporting that satisfies regulatory scrutiny without creating unnecessary operational overhead. If your board cannot currently answer the AI cybersecurity readiness questions in the checklist above, that is where we start.

For organizations that need fractional CISO leadership to operationalize these programs, my vCISO service provides senior security leadership without the full-time executive overhead. If your challenge is compliance program maturity across multiple overlapping frameworks, my compliance advisory service maps your controls once and satisfies many. And for boards that need direct AI and cybersecurity education delivered in the boardroom, my board advisory service gives directors the fluency they need to ask the right questions and recognize credible answers.

The Cyber AI Profile's final version will arrive in 2026 with considerably more regulatory weight behind it. The organizations that begin alignment now will spend that time building capability. Those that wait will spend it catching up under audit pressure.

Schedule a discovery conversation to assess where your AI cybersecurity governance stands and what it will take to reach a defensible posture before the deadline matters.