AI Inventory: The Three Categories Most Institutions Miss

An AI inventory is the artifact every AI governance framework depends on and the one most institutions discover is incomplete only after the governance work has already started. The reason is not negligence. It is that inventories built by asking teams what AI they use return the systems people think of as AI, and a growing share of the AI inside a financial institution arrived attached to something else entirely. This article covers what an AI inventory has to capture, the three categories it needs to cover rather than one, and why the inventory is the gating artifact for scoping an AI governance program rather than a housekeeping exercise to complete afterward. Why the AI Inventory Gates Everything Downstream Every Framework Assumes It Exists The AI governance frameworks a US financial institution is likely to encounter all assume a complete inventory before their own requirements make sense. The NIST AI RMF Map function asks an organization to establish the context and categorize its AI use cases, which is not possible against an unknown population. ISO/IEC 42001 expects identification of the AI systems within the scope of the management system. The Financial Services AI Risk Management Framework published by the Cyber Risk Institute uses an adoption stage questionnaire whose answers depend directly on what AI the institution actually runs and what that AI touches. None of those frameworks is unusual in this. The pattern repeats because risk work cannot be scoped against an unenumerated population, and every one of these frameworks is fundamentally a scoping exercise before it is a control exercise. An Incomplete Inventory Produces a Wrong Stage For an institution working toward the FS AI RMF specifically, the inventory does more than inform the program. It determines the size of the program. The adoption stage questionnaire places an institution in one of four stages based on business impact, technology implementation, and scalability, and the stage determines which control objectives apply. An institution assessed at Minimal carries 120 control objectives. An institution assessed at Evolving carries 193. The line between those two stages is crossed the moment AI touches sensitive data or an external-facing outcome. That is a low threshold, and it is exactly the kind of threshold an incomplete inventory hides. An institution that catalogued only the models its data science team built, and missed a vendor product whose AI feature reaches customer communications, will assess itself at a stage below its actual footprint and scope a program 73 control objectives too small. The error surfaces later, in front of Internal Audit or an examiner, at the point where it is most expensive to correct. The Failure Mode Is Predictable Institutions do not usually fail at this by skipping the inventory. They fail by building one from the two sources that feel authoritative and are not: the model inventory the model risk function already maintains, and a survey sent to business unit leads. The model inventory is incomplete because it was built to catalogue models, which is a narrower object than AI capability. It was scoped around systems that produce a quantitative output someone validates, which excludes a summarization feature, a drafting assistant, or a classification step inside a workflow tool. Those are not models in the sense the inventory was designed around, and they were correctly excluded under the definition that inventory was built to serve. The survey is incomplete for a different reason. It returns what people classify as AI, and the systems that create the most governance exposure are frequently the ones nobody classifies that way. A relationship manager using a platform feature that drafts client summaries does not experience that as using AI, and will not report it as such. Both sources are worth using. Neither is sufficient alone, and an inventory built from both together still misses an entire category. What an AI Asset Inventory Has to Capture Three Categories, Not One The correction is structural rather than procedural. An inventory scoped for AI governance covers three categories of system, and most institutions have a reliable process for only the first. Category What it covers How it is usually found Built Models and AI systems the institution developed internally Existing model inventory, data science team records Procured AI capabilities the institution knowingly bought as AI Contracts, procurement records, vendor risk files Embedded AI that arrived inside products bought for other reasons Product release notes, configuration review, technical discovery The third row is where the real work sits. A productivity suite that added a drafting assistant, a service platform that added ticket summarization, a core banking vendor that added anomaly narration, and a CRM that added lead scoring all introduced AI into the institution without a single procurement decision that named AI as the thing being purchased. No internal record identifies those capabilities as AI, because at the time of purchase they were not the thing being evaluated. Embedded AI also tends to be the category with the shortest path to a customer-facing outcome, which means it is disproportionately likely to be what pushes an institution across a stage boundary. That combination, high governance impact and low discoverability, is what makes it worth a dedicated discovery effort rather than another survey question. What Each Entry Needs to Carry An inventory that lists systems without attributes is a list, not an inventory. Each entry needs enough structure to support the risk classification and stage determination that follow. Attribute Why it is needed Named owner Assigns accountability and gives the control set a person to route to Use case Determines whether the system touches a critical function or decision Data classification Determines whether the system reaches sensitive or regulated data Externally facing Determines whether outputs reach customers or counterparties Build, procured, or embedded Determines which controls the institution runs and which it inherits AI type Separates traditional statistical models from generative and agentic systems The last row carries more weight in 2026 than it did a year earlier. The revised interagency model risk management guidance issued

Model Risk Management Framework: What to Rebuild After 2026

A model risk management framework written before April 2026 now rests on guidance that no longer exists. The rescission of SR 11-7 and its companion issuances did not simply retire a citation. It narrowed the definition of what counts as a model, removed the prescriptive methodology that validation programs were built around, introduced an asset-size marker for relevance, and folded a specialized BSA and AML regime back into the general framework without replacement. Each of those changes alters a different part of the framework document, and an institution that updates only the citations will carry a policy that describes a program the guidance no longer asks for. This article covers what each change means structurally and what a framework rebuild actually has to touch. What the Revised Guidance Actually Says Three Areas, Not a Methodology The revised guidance, issued April 17, 2026 as OCC Bulletin 2026-13, Federal Reserve SR 26-2, and FDIC FIL-15-2026, organizes its content around the factors that influence model risk and the features of effective practice in three areas: model development and model use, including model testing; model validation and monitoring, including validation of conceptual soundness and outcomes analyses; and governance and controls, including clear policies and roles and responsibilities. It adds a dedicated treatment of vendor and other third-party products, including validation of those products. Read against the 2011 framework, the continuity is real and the change is in specificity rather than subject matter. Conceptual soundness and outcomes analysis survive as validation concepts. What does not survive is the detailed operating model built on top of them: prescribed validation frequencies, enumerated board and internal audit task lists, and the specific methodology requirements that most framework documents reproduced almost verbatim from the guidance. What Survives the Change Unchanged Before cataloguing what moved, it is worth being precise about what did not, because a rebuild that rewrites settled sections wastes effort and introduces error. Conceptual soundness remains a validation concept. Outcomes analysis remains a validation concept. Ongoing monitoring remains an expectation. Clear policies and documented roles and responsibilities remain a governance expectation. Effective challenge by parties independent of model development and use remains the mechanism by which validation carries weight. An institution whose framework describes those things well already holds most of a compliant document. What it does not hold is the justification layer underneath them, which the 2011 guidance supplied and the revised guidance withdrew. The rebuild adds reasoning to sections that survive and rewrites the sections that did not. The Enforceability Statement Changes the Drafting Problem The guidance states that it does not set forth enforceable standards or prescriptive requirements and that non-compliance will not result in supervisory criticism. That sentence is usually reported as a relaxation, and it functions as a drafting burden. Under prescriptive guidance, a framework document could justify its choices by citing the source. A validation cycle was annual because the guidance implied annual. Under principles-based guidance that disclaims enforceability, the same document has to justify the same choice on the institution’s own risk rationale, and a reviewer asking why the cycle is annual will not accept a citation as the answer. The framework now has to carry the reasoning it previously borrowed. The Relevance Marker at $30 Billion The OCC states that the updated guidance is expected to be most relevant to banking organizations with over $30 billion in total assets, while allowing relevance for certain smaller institutions. This is a marker rather than a scoping rule, and it should be read carefully in both directions. For an institution above the threshold, nothing about the marker reduces expectations. For an institution below it, the marker does not mean model risk management stops applying. It means the depth and formality of the program should be commensurate with the institution’s model use rather than modeled on what a large institution runs, which is a point the agencies had made in earlier clarifications and have now placed in the guidance itself. The practical effect for a smaller institution is that a framework copied from a larger peer is now harder to defend as a proportionate choice, not easier. The Narrower Model Definition Changes the Inventory What Left the Definition The revised guidance narrows what counts as a model to complex quantitative methods, systems, or approaches applying statistical, economic, or financial theories. Re-scoping against it is an AI and model inventory exercise before it is a policy exercise. Several categories of tool that many institutions had pulled into model inventory over the previous decade fall outside that narrower definition, including deterministic rule-based tools and spreadsheet calculations that perform no statistical estimation, alongside the explicit exclusion of generative and agentic AI. Item Status under the revised definition Statistical and econometric models In scope Machine learning models producing quantitative estimates Generally in scope Deterministic rule-based tools Outside the narrower definition Spreadsheet calculations without statistical estimation Outside the narrower definition Generative AI and agentic AI Explicitly excluded, pending a forthcoming request for information The table describes what the guidance reaches, not what the institution can stop caring about, and conflating the two is the most expensive available mistake here. A rule-based sanctions screening tool that falls outside the model definition still produces regulatory outcomes the institution answers for. What changed is which framework governs it, not whether it needs governing. Removing Items Requires a Documented Rationale The temptation after a narrowing definition is to trim the inventory, and trimming is legitimate. Doing it silently is not. Every item removed from model inventory needs a recorded rationale explaining why it falls outside the revised definition, and a statement of what governs it instead. That second half is the part institutions skip. An item removed from model inventory with no successor control regime named has not been reclassified. It has been dropped, and an examiner reading an inventory that shrank by forty percent in one cycle will ask where those items went. A rebuild that produces a smaller inventory and a companion register of reclassified items with their new

AI Red Teaming for AI Agents: How to Test What Your Agent Does, Not Only What It Says

AI red teaming for AI agents means attacking an agent the way an adversary would: through its prompts, the content it retrieves, the tools it can call and the permissions it holds. The goal is to learn what the agent actually does under pressure. The stakes changed when models started taking actions. In 2025, a single crafted email was enough to make an enterprise copilot exfiltrate internal files without anyone clicking anything, a flaw rated 9.3 on the CVSS scale. This guide covers what agent red teaming must include, how it differs from a penetration test or a model evaluation, how certification testing is reshaping it, and how to build a program that produces evidence your buyers and auditors will accept. Why AI Red Teaming Has Moved From Research Labs to the Security Review Agents turned output risk into action risk A language model produces text that a person usually reads before anything happens. An agent calls an API, updates a record, sends a message or runs code, often with no human reviewing each step. That shift moves the risk from what the system says to what the system does, and it changes what a test has to prove. A jailbreak that makes a chatbot say something embarrassing is a reputational problem. The same manipulation applied to an agent with write access to a CRM, a payments API or a code repository is a security incident. Buyers now ask for behavioral evidence Enterprise security reviews used to accept policies, architecture diagrams and a SOC 2 report as proof that a product was safe to deploy. For agents, that evidence describes intended boundaries without showing whether the agent respects them when someone pushes past them. AIUC-1 is a certification standard written specifically for AI agents. It closes that gap by pairing an audit of controls with large-scale technical testing of the agent itself, and it produces an audit report that includes red teaming results buyers can request. The standard now sits in the Cloud Security Alliance STAR Registry, and vendors including UiPath, Cursor, Harvey, ElevenLabs and Fin have certified. Whether or not you pursue certification, those announcements have taught procurement teams what a credible answer to “how do you know your agent is safe?” looks like. The cost of a finding someone else discovers A failure your own testing finds is a ticket. The same failure found by a customer’s security team during a review is a stalled deal, and found by an attacker in production it is a disclosure event. Agent vulnerabilities are also unusually cheap to exploit. The attack is written in natural language and often delivered through content the agent reads on its own, such as an email, a support ticket, a web page or a shared document. Elevate’s analysis of agentic applications under attack shows how memory, tools and external content multiply the entry points. The economics favor finding these failures first, on your schedule. How AI Red Teaming Differs From Penetration Testing and Model Evaluation Three disciplines get grouped under “AI security testing,” and they answer different questions. Confusing them is the most common reason a test report fails to satisfy the reviewer who asked for it. Each one is useful, but only one of them tests the behavior of an agent operating inside your environment. Dimension Application penetration test Model red teaming AI agent red teaming Target Infrastructure, APIs, authentication The model’s responses The agent’s decisions and actions across tools, data and channels Core question Can an attacker break the system around the AI? Can the model be made to say something harmful? Can the agent be made to do something it should refuse? Typical findings Misconfigurations, injection flaws, broken access control Jailbreaks, toxic output, sensitive data in responses Goal hijack, tool misuse, privilege abuse, unsafe tool calls Evidence produced Findings with CVSS severity Pass rates on prompt sets Scenario results tied to controls, severity thresholds and retest history The three columns complement each other; none replaces the others. A penetration test can confirm that the API gateway enforces authentication while the agent behind it executes an instruction hidden in a customer email. Model red teaming can show that the underlying model refuses harmful requests in isolation. Yet the same model, wrapped in an agent with broad tool permissions, can be walked into a harmful action one reasonable-looking step at a time. Agent red teaming builds on both, and it answers the question an enterprise buyer actually cares about. Elevate’s overview of the OWASP LLM Top 10 covers the model-level risks this layer inherits. The attack surface is everything the agent can reach An agent’s exposure equals the sum of the credentials, tools, data sources and other agents it can touch. That is why scoping matters more in agent testing than in most security work. An agent with read-only access to a knowledge base and an agent that can issue refunds are different risk classes, even when they run on the same model. A useful scope document lists every tool the agent can call, the permissions attached to each, the content sources it ingests without human review, and the channels through which users reach it. If that list does not exist, the red team will spend its first week building it, and that is worth knowing before you budget the engagement. Non-deterministic behavior changes what a pass means Traditional tests are largely repeatable: an injection flaw either exists or it does not. An agent can refuse an attack nine times and comply on the tenth, depending on phrasing, context length or the state of its memory. Agent red teaming therefore runs large scenario sets, repeats them, and measures failure rates by severity instead of declaring a single pass. Certification testing reflects this: AIUC-1 evaluations typically run 1,000 to 5,000 test scenarios, and Harvey reported passing more than 3,000 unique tests with zero critical failures. Define in advance which failure severities are acceptable at what rate, or the results will be impossible to interpret. What

AI Governance Maturity Model: Why Self-Scored Tiers Fail

An AI governance maturity model is supposed to tell an organization how much governance it needs, and most of them fail at exactly that because they let the organization answer the question about itself. A model that asks a team to rate its own AI governance on a scale of one to five returns a number shaped by how the team feels about its program rather than by what its AI actually does. The models that work invert this: they assess observable facts about the AI footprint and derive the maturity level from those facts, which means the organization can disagree with the result and still be bound by it. This article covers what separates a descriptive maturity model from a determinative one, how the financial services sector built the latter, and what happens when an institution assesses itself below its real position. What an AI Governance Maturity Model Is For The Purpose Is Scoping, Not Scoring The common assumption is that a maturity model exists to measure progress, with an organization moving from level one to level five as its program improves. That framing works for a capability model where the levels describe how well something is done. It works poorly for AI governance, where the question is not how well the organization governs but how much governance the organization’s AI use actually requires. Those are different questions with different answers. An organization running a single internal chatbot with excellent documentation and clear ownership is doing governance well and needs very little of it. An organization with AI embedded across customer-facing decisions and no documentation at all is doing governance badly and needs a great deal. A model that scores quality tells the first organization it is mature and the second that it is not, which is true and useless, because neither answer tells either organization what control set to build. A maturity model built for scoping answers the useful question instead. It establishes what the AI footprint is, and the obligations follow from that rather than from an assessment of how well the organization currently handles them. Why Uniform Control Sets Do Not Work The alternative to a maturity model is applying the same control set to every organization, and that alternative fails in both directions at once. A control set sized for an institution running autonomous decisioning in critical functions is unusable at a community bank with two AI pilots. A control set sized for the community bank leaves the larger institution materially under-governed. This is the specific problem a staged model solves, and it is why sector frameworks have converged on staging rather than on a single baseline. The value is not in the label an organization receives. It is in the fact that the label determines a scope, which makes the same framework document usable across institution sizes that otherwise have nothing operationally in common. Why Self-Assessment Produces the Wrong Level Organizations Estimate From Impression When an organization places itself on a maturity scale without a determinative methodology, the answer reflects the felt state of the program rather than the actual footprint. Teams that know their documentation is thin, their ownership unclear, and their monitoring informal will place themselves low, because low is what the program feels like from inside. The felt state and the required scope are unrelated. An organization can have genuinely immature governance practices and an AI footprint that demands a large control set, and in fact that combination is the most common one, because footprints grow faster than governance programs. Self-assessment produces a level that describes the program and is then used to size the program, which is circular in a way that consistently under-scopes. The Averaging Problem The second failure is subtler and more common. A model assessed across several dimensions invites an organization to average across them, and averaging is almost always wrong for risk. An institution might look modest on technology implementation, having built nothing internally, modest on scalability, with AI confined to a few teams, and significant on business impact, because one of those few deployments touches customer communications. Averaged, that reads as a low or middle position. In risk terms it is not, because the exposure comes from the dimension with the highest reading rather than from the mean. The customer-facing deployment carries its consequences regardless of how little else the institution runs. A determinative model removes the option to average. It resolves to the highest position the organization matches on any dimension, which produces a result the organization frequently disagrees with and which is defensible precisely because the organization could not talk itself out of it. How a Determinative Model Works The Financial Services Example The clearest working example of a determinative AI governance maturity model in production is the adoption stage methodology inside the Financial Services AI Risk Management Framework. The Cyber Risk Institute published it at version 1.0 in February 2026, developed with the Financial Services Sector Coordinating Council, the US Department of the Treasury, and more than 100 financial institutions. It is a voluntary industry framework rather than a supervisory rule, and its weight comes from that sector consensus rather than from an agency issuing it. The methodology places an institution in one of four stages, assessed across three dimensions: business impact, technology implementation, and scalability. What makes it determinative rather than descriptive is the mechanic underneath. The questionnaire runs from the highest stage downward and stops at the first stage with a matching statement, which means an institution lands at the highest stage it matches anywhere rather than at an average across the three dimensions. The Four Stages Stage What characterizes it Initial AI is not embedded in critical functions or business decisions. Predictive AI on legacy systems, no modern AI, nothing external-facing, no internal model development Minimal Limited use for non-critical tasks. Narrow, not external-facing, no sensitive data, no internal model development, no infrastructure to scale Evolving Drives outcomes through external-facing solutions and sensitive data, but not critical decision-making.

SR 11-7 Was Rescinded, and the Guidance That Replaced It Excludes AI

SR 11-7 no longer governs model risk management at US banking organizations. On April 17, 2026, the Office of the Comptroller of the Currency, the Federal Reserve Board, and the Federal Deposit Insurance Corporation jointly issued revised interagency guidance that rescinds the 2011 framework outright, and the document that replaced it carries a scope exclusion most institutions have not yet worked through: generative AI and agentic AI models sit outside it. For a risk function that spent fifteen years building validation programs against SR 11-7, the practical question is no longer how to extend that framework to cover AI. It is what governs the AI that the new guidance declined to reach. What Changed on April 17, 2026 The Rescission Was Total, Not a Revision The agencies did not amend SR 11-7. They replaced it and withdrew the surrounding body of guidance with it. OCC Bulletin 2026-13, issued alongside Federal Reserve SR 26-2, rescinds OCC Bulletin 2011-12, the OCC companion to SR 11-7, along with three other issuances that together formed the operating manual most model risk functions were built from. Rescinded issuance What it covered OCC Bulletin 2011-12 / Fed SR 11-7 The foundational supervisory guidance on model risk management OCC Bulletin 2021-19 / Fed SR 21-8 The interagency statement on model risk for BSA and AML systems OCC Bulletin 1997-24 Examination guidance on credit scoring models Model Risk Management booklet, Comptroller’s Handbook The examination procedures built on the 2011 guidance Reading that list together rather than item by item is what makes the scale clear. An institution that lost only SR 11-7 would still have the BSA and AML statement and the Handbook booklet to anchor its validation methodology. Losing all four in a single issuance removes the accumulated interpretation as well as the source document, which is why a model risk function cannot treat this as a citation update in its policy library. The New Guidance Is Voluntary in a Way SR 11-7 Was Not The more consequential change is the enforcement posture. SR 11-7 was technically supervisory guidance rather than a rule, but examiners applied it as a baseline expectation and cited deficiencies against it in supervisory findings. The 2026 guidance states plainly that it does not set forth enforceable standards or prescriptive requirements, and that non-compliance will not result in supervisory criticism. That sentence reads like relief and functions like the opposite. A prescriptive standard tells an institution what adequate looks like, which means an institution that meets it has a defensible position. Principles-based guidance that disclaims enforceability transfers the definition of adequate from the supervisor to the institution, and an institution that has to define adequate for itself needs a documented methodology behind that definition. The supervisory expectation did not disappear. The specification of how to meet it did. What the Revised Guidance Keeps The revised guidance retains the conceptual core of validation, including conceptual soundness, outcomes analysis, and ongoing monitoring, while removing the prescriptive methodology around them. Traditional statistical and machine learning models used in credit underwriting, fraud detection, and transaction monitoring generally remain within scope. An institution running a logistic regression credit scorecard or a gradient boosting fraud model is still doing model risk management, and the 2026 principles still describe what the agencies expect from it. The change for those models is a shift from a detailed operating model to an outcomes description. A validation program built around fixed cycles, prescribed backtesting frequencies, and specific board task lists now has to justify those choices on its own risk rationale rather than by pointing to a paragraph in the 2011 guidance. The Scope Exclusion That Creates the Gap Generative and Agentic AI Are Outside the Guidance The revised guidance excludes generative AI and agentic AI models from its scope, on the stated basis that these technologies are novel and rapidly evolving. The agencies also announced that they plan to issue a request for information in the near future addressing model risk management generally and banks’ use of AI specifically, including generative AI, agentic AI, and AI-based models. Two things follow from that, and institutions tend to get one of them right and the other wrong. The first is widely understood and rarely acted on; the second is widely misunderstood and expensive. The first is straightforward: the exclusion is a deferral, not an exemption. Nothing about the underlying activity changed. A generative AI system that drafts customer communications, summarizes credit files, or routes service requests still produces outcomes the institution is accountable for, still touches consumer data, and still sits inside the obligations attached to that activity. The agencies withdrew the specification of what adequate governance looks like for these systems. They did not withdraw the accountability. The second is where programs go wrong: the exclusion is not a reason to wait for the request for information. An RFI has no published timeline, and a request for information is a step before a proposal, which is itself a step before a final issuance. Counting those steps against how fast generative AI is being deployed inside institutions produces an uncomfortable answer about how long the interval will run. An institution deploying generative AI across its operations during that interval is accumulating a governance record that a future examiner will read retrospectively. The examination will not ask whether guidance existed at the time. It will ask what the institution did about a known risk during a period when the supervisors had publicly flagged that they intended to address it. Waiting produces an unexplainable gap in that record rather than a clean slate, and the institutions that fare worst are usually the ones that read the exclusion as permission and documented nothing for eighteen months. What the Exclusion Actually Removes It helps to be specific about what an institution loses when a system falls outside the revised guidance, because the answer is narrower than it first appears and more awkward than a simple gap. The exclusion removes one layer of a stack rather than the

AI Governance for Procurement and Vendor Reviews

An AI vendor risk assessment run by a procurement team that treats it like any other software vendor questionnaire will miss almost everything that actually matters about the vendor’s AI system. Standard vendor due diligence asks about data security, uptime, and financial stability. It rarely asks what the AI model was trained on, how the vendor handles a biased or hallucinated output, or what happens when the vendor updates the underlying model without notice. Procurement teams evaluating AI vendors need a genuinely different review process, not the existing vendor questionnaire with a few AI-flavored questions added at the end. This article covers what an AI-specific vendor review needs to ask, how to structure the process so it does not stall procurement, and where AI governance frameworks like ISO 42001 fit into a review most organizations run without ever pursuing certification themselves. Why Standard Vendor Due Diligence Falls Short for AI A traditional vendor security questionnaire assumes a relatively stable product: the vendor built something, and unless a major update occurs, its behavior stays consistent. An AI vendor, particularly one built on a third-party foundation model, does not fit this assumption. The underlying model can change on a schedule the vendor does not fully control, the outputs can vary in ways traditional software does not, and the vendor’s own supply chain often includes AI components from other companies the procuring organization has never heard of and cannot directly assess. The Questions a Standard Questionnaire Never Asks A standard vendor review asks about data encryption, access controls, and incident response, all of which remain relevant for an AI vendor. What it typically omits entirely is any question about the AI system’s training data provenance, the vendor’s process for monitoring output quality after deployment, how the vendor handles a discovered bias in the model’s outputs, and what recourse the procuring organization has if the vendor’s underlying model provider changes terms or discontinues the model version the organization built its integration around. These gaps are not edge cases. They are the questions that determine whether an AI vendor relationship is defensible a year into the contract. Why This Matters More for AI Than for Traditional Software A traditional software vendor’s product behaves the same way today as it did last year unless the vendor ships a new release the customer can review and test before adopting. An AI vendor’s product can behave differently for reasons entirely outside the customer’s visibility: a foundation model provider retrains or deprecates a model version, a data source feeding the system changes in composition, or the vendor tunes its own system prompt or fine-tuning approach without treating it as a customer-facing release at all. This is the structural reason AI vendor risk cannot be evaluated once at procurement and then set aside. The product the organization is buying is less stable, by nature, than the traditional software it is used to vetting. What an AI Vendor Risk Assessment Needs to Cover A rigorous AI vendor review organizes around four areas that a standard security questionnaire does not reach. Model Provenance and Training Data The vendor should be able to describe, at least at a summary level, what data trained the model underlying its product, whether that model is proprietary or licensed from a third-party foundation model provider, and what license or usage terms govern that underlying relationship. An organization cannot fully evaluate its own exposure to a vendor’s AI risk without understanding whether that vendor is the actual model developer or a reseller of someone else’s model wrapped in a product interface, since the two carry very different risk profiles and very different points of failure. Output Monitoring and Quality Control Ask the vendor directly how it monitors its own AI system’s outputs for accuracy, bias, or degradation after deployment, not just during initial development and testing. A vendor with a mature AI governance program can describe a specific, ongoing monitoring process. A vendor that treats initial testing as sufficient and has no ongoing output quality process is asking the procuring organization to absorb model drift risk it has no visibility into. Change Management and Model Updates AI models change in ways traditional software does not always signal clearly. Ask how the vendor notifies customers of underlying model changes, whether customers can control or delay adopting a new model version, and what testing the vendor performs before deploying a model update to production. A vendor that pushes model updates silently, with no customer notification or opt-out mechanism, transfers real risk to every customer relying on consistent behavior from that system. Third-Party AI Supply Chain Most AI vendors are not building foundation models from scratch. They are building on top of models from a small number of large providers, sometimes layering multiple third-party AI components together. Understanding this chain, and confirming the vendor has visibility into and accountability for the components it builds on, prevents a procuring organization from discovering during an incident that its vendor’s AI risk actually traces back to a sub-vendor nobody in the review process ever identified. A Note on Public Sector Procurement Government agencies procuring AI face a related but distinct set of considerations beyond the vendor-facing review this article describes. The basics of ethical AI procurement for government buyers covers obligations that apply specifically when the procuring entity is itself a public body, including transparency and public accountability requirements that a private-sector vendor review does not need to address in the same way. A commercial organization selling into government, rather than a government agency itself procuring, should understand both sides of this relationship: its own vendor review process for the AI components it builds on, and the ethical procurement standard the government customer on the other side of the table is likely applying to it. Building the Review Into the Procurement Process An AI-specific review only works if it is built into the procurement workflow at the right stage, rather than bolted on after a vendor has already been selected on other criteria.

EU AI Act Article 50: What It Actually Requires Now

EU AI Act Article 50 is already in force. It took effect on August 2, 2026, and the widely reported delay to the Act’s high-risk obligations did not touch it. Many organizations read the headlines about the EU’s Digital Omnibus pushing back the Act’s most demanding requirements, filed the entire regulation under next year’s problem, and moved on. That assumption is now a compliance gap, not a scheduling one, because Article 50’s transparency obligations apply today, regardless of whether an organization’s AI systems are classified as high-risk under the Act’s separate Annex III framework. This article explains what changed, what Article 50 actually requires, why it reaches organizations that do not think of themselves as AI companies, and how to fold compliance into governance infrastructure many regulated organizations are already building. Article 50 Is Already in Force, and the Omnibus Did Not Change That The confusion here is understandable, because the EU AI Act’s implementation timeline genuinely did change in 2026, just not in the way most coverage suggested. What the Digital Omnibus Actually Delayed The EU’s Digital Omnibus on AI, formally Regulation (EU) 2026/1744, entered into force on July 27, 2026, after final adoption on June 29, 2026. It pushed the compliance deadline for high-risk AI systems under Annex III from August 2, 2026 to December 2, 2027, and delayed the Annex I product-safety obligations to August 2, 2028. Those are real, substantial delays, and they are the reason so much of the public discussion around the EU AI Act in mid-2026 described the regulation as having been pushed back by more than a year. What that coverage largely missed is that the Omnibus explicitly left Article 50 untouched. The transparency obligations in Article 50 became generally applicable and enforceable by national authorities across the EU on August 2, 2026, on schedule, and they apply regardless of whether the underlying AI system is classified as high-risk. An organization that correctly tracked the Annex III delay and incorrectly assumed the same delay covered Article 50 is not behind schedule. It is currently out of compliance with a regulation that has already taken legal effect. The One Narrow Exception The Omnibus did make one specific, limited change relevant to Article 50. Providers of AI systems that generate synthetic audio, image, video, or text content, and that placed those systems on the market before August 2, 2026, received a four-month grace period on the machine-readable marking requirement under Article 50(2), extending that specific deadline to December 2, 2026. This exception is narrow in three ways: it applies only to the marking obligation, only to systems already on the market before the original deadline, and not at all to the other three transparency obligations Article 50 establishes, all of which remain in force from August 2, 2026 with no grace period. What Non-Compliance Actually Costs Article 99 of the EU AI Act sets the enforcement framework, and it assigns Article 50 breaches to the middle of the Act’s three penalty tiers. Non-compliance with Article 50’s transparency obligations can result in administrative fines of up to fifteen million euros or three percent of an organization’s total worldwide annual turnover for the preceding financial year, whichever figure is higher. This sits below the top tier reserved for the Act’s prohibited practices under Article 5, which reaches thirty-five million euros or seven percent of turnover, but it is a real, material exposure rather than a symbolic one. The Regulation caps the fine at the lower of the two figures for small and medium enterprises, and enforcement sits with each member state’s national market surveillance authority, which took on that enforcement power on the same August 2, 2026 date the obligations themselves became applicable. Reviewing the full EU AI Act compliance timeline puts this specific enforcement date in context alongside the other deadlines the Digital Omnibus did and did not move. The Four Transparency Obligations Under Article 50 Article 50 covers four distinct situations, each triggering a specific disclosure duty. Understanding which apply to a given organization requires looking at actual AI use cases, not at industry classification. AI Systems That Interact Directly With People Providers of AI systems intended to interact directly with natural persons must ensure those individuals are informed they are interacting with an AI system, unless it is obvious from the circumstances to a reasonably well-informed person. This obligation covers chatbots, virtual assistants, and any conversational AI interface a person might otherwise reasonably mistake for a human. The disclosure requirement is not satisfied by a policy buried in terms of service. It requires the interaction itself to make the AI nature of the system reasonably clear to the person engaging with it. Marking AI-Generated Content Providers of AI systems that generate synthetic audio, image, video, or text content must mark that output in a machine-readable format detectable as artificially generated or manipulated. This obligation is covered in depth in Elevate’s guide to labeling AI-generated content under the EU AI Act, including the specific format requirements and the narrow grace period for legacy systems described above. Emotion Recognition and Biometric Categorization Disclosure Deployers of an emotion recognition system or a biometric categorization system must inform the individuals exposed to that system of its operation, and must process any personal data involved in line with EU data protection law. An exception applies to systems used to detect, prevent, or investigate criminal offenses, subject to safeguards under other applicable law. This obligation attaches to the deployer, the organization using the system, not only to the technology’s original provider, which means an organization that purchases and operates a third-party emotion-recognition or biometric-categorization tool carries this disclosure duty directly. Deepfake Disclosure Deployers of AI systems that generate or manipulate image, audio, or video content constituting a deepfake must disclose that the content has been artificially generated or manipulated. This obligation sits alongside the provider-side marking requirement under Article 50(2) but applies specifically to the deployer’s disclosure to the audience encountering the content, which is a distinct duty

EU AI Act Compliance Checklist: Obligations by Risk Category

An EU AI Act compliance checklist only works if it is organized the way the law itself is organized: by risk tier. The Act does not impose one uniform set of rules on every AI system; it sorts systems into risk categories and assigns obligations to each, so the first question is never what must be done but rather which tier a system falls into. This guide lays out the obligations for the prohibited, high-risk, and limited-risk tiers, explains how classification drives everything, and points to where the current application dates are maintained, since those have been revised. The reason the tier comes first is that it determines the entire obligation set. A system in the minimal-risk tier carries no mandatory obligations, while one in the high-risk tier carries an extensive program of them, and misclassifying a system is therefore the most consequential mistake an organization can make. Working through this checklist means classifying each AI system first, then applying only the obligations that its tier actually triggers. How the EU AI Act Classifies Risk The Act’s risk-based structure is the foundation of every obligation, and the table sets out the four tiers and what each demands. The classification comes directly from the Regulation. Risk tier What it covers What the Act requires Prohibited AI practices deemed an unacceptable risk under Article 5, such as certain manipulative techniques and social scoring The practice is banned outright High-risk AI that is a product safety component or falls within the Act’s listed use cases under Article 6 An extensive set of obligations before and throughout use Limited risk AI that interacts directly with people or generates content, under Article 50 Transparency obligations Minimal risk All other AI systems No mandatory obligations under the Act The practical consequence of this structure is that classification is not a formality but the step that decides how much work an organization faces. Most business AI falls into either the high-risk or limited-risk tiers, and the gap between the two is large, so an honest classification, rather than an optimistic one, is where compliance genuinely begins. The tiers are mutually exclusive for a given use, which is why the checklist is built to be read one tier at a time. Prohibited Practices: What You Must Not Do The prohibited tier is the shortest part of any such checklist because it is a list of things to stop, not build. Article 5 bans a defined set of practices outright, including AI that uses subliminal or purposefully manipulative techniques to materially distort behavior, AI that exploits the vulnerabilities of specific groups, social scoring of individuals by public or private actors, and certain uses of biometric categorization and remote identification. An organization’s obligation here is simply to confirm that none of its systems, current or planned, fall into these categories. The reason this tier sits first on the checklist is that a prohibited practice cannot be remediated into compliance; it must be abandoned. Unlike a high-risk system, which becomes compliant through controls and documentation, a system whose purpose is a prohibited practice has no compliant version. Screening the AI portfolio against Article 5 early therefore prevents an organization from investing in a system it can never lawfully deploy. High-Risk Systems: The Core Obligations The high-risk tier carries the heaviest obligations, and it is where most of the checklist lives. A system is high-risk under Article 6 either when it is a safety component of a regulated product or when it falls within the Act’s listed high-risk use cases, which cover sensitive domains such as employment, access to essential services, and critical infrastructure. Once a system is classified high-risk, a defined program of requirements applies to it before it reaches the market and throughout its operation. The core requirements the Regulation sets for high-risk systems are a connected set, and a checklist should treat them as items to evidence rather than boxes to tick. They include a risk management system that runs across the lifecycle under Article 9, data and data governance practices under Article 10, technical documentation under Article 11, automatic record-keeping and logging under Article 12, transparency and information provided to deployers under Article 13, human oversight designed into the system under Article 14, and appropriate accuracy, robustness, and cybersecurity under Article 15. Surrounding these, providers must operate a quality management system, complete a conformity assessment before placing the system on the market, register it as required, and monitor it after deployment. The through-line across these obligations is evidence: each requirement is satisfied not by intent but by documentation that a regulator or a deployer can inspect, which is why a high-risk program looks in practice like a managed system rather than a one-time project. This is also where a structured AI risk assessment does much of the work, because the risk management requirement sits at the center of the high-risk obligations and feeds several of the others. Limited-Risk Systems: Transparency Obligations The limited-risk tier is defined by transparency rather than by the extensive controls of the high-risk tier. Under Article 50, systems that interact directly with people must make clear that a person is dealing with an AI system unless that is already obvious, and systems that generate or manipulate content must ensure that synthetic content, including deepfakes, is disclosed and detectable. Similar disclosure obligations apply to systems that perform emotion recognition or biometric categorization, which must inform the people exposed to them. The obligation for this tier is narrower but still real, and organizations often overlook it precisely because it feels light. A chatbot, a synthetic-media tool, or a recommendation interface may carry no high-risk obligations yet still owe clear disclosure to the people who encounter it. The checklist item for this tier is therefore to identify every system that interacts with people or produces content and confirm that its disclosures meet the transparency standard. Who the Obligations Fall On A checklist organized by risk tier answers what the obligations are, but a second question decides

AI Risk Assessment Template: The Fields Auditors Look For

An AI risk assessment template is only as useful as the fields it captures, and the fields that matter are the ones an auditor or regulator expects to see: not just the technical risk, but its impact on the people the AI system affects. A template that treats an AI risk like any other IT risk misses the element that AI governance frameworks specifically require, which is the assessment of impact on individuals and society. This guide sets out the fields an AI risk assessment template should contain, a scoring approach that prioritizes correctly, and a structure aligned to ISO 42001 and the NIST AI Risk Management Framework. The reason to build the assessment around the right fields, rather than reach for a generic risk register, is that AI risk has a dimension ordinary risk assessments do not capture. An AI system can create risks to fairness, safety, and the rights of the people it affects, and the frameworks that govern AI expect those to be assessed explicitly. A template that prompts for them turns an AI risk assessment from a repurposed security exercise into one that actually satisfies what AI governance requires. What an AI Risk Assessment Template Should Capture A strong template captures each risk in enough structured detail to assess, prioritize, and act on it, without becoming so heavy that no one completes it. At minimum it should identify the AI system or use case being assessed, describe the context and purpose, state the specific risk, and capture the AI-specific dimension of how that risk could affect individuals. It then records the likelihood and severity that produce a rating, the controls already in place, the resulting risk level, and the treatment plan with an owner. Each field does a distinct job, and the assessment is only as good as its weakest field. The field that distinguishes an AI risk assessment from a generic one is the impact on individuals and society. Ordinary risk assessments focus on impact to the organization, but AI governance frameworks require an organization to consider how an AI system affects the people subject to it, which is the substance of an AI impact assessment. Building that field into the template rather than bolting it on afterward is what makes the assessment defensible under AI governance scrutiny. The Fields That Matter The table sets out the fields a defensible template should contain, what each captures, and why it matters. It is a structure to adapt to the organization’s AI systems, not a fixed form. Field What it captures Why it matters AI system or use case The specific system or application being assessed Scopes the assessment to one clear subject Context and purpose What the system does, for whom, and in what setting Frames how the risk should be judged Risk description The specific risk, from security to fairness to safety The core of the entry Impact on individuals How the risk could affect the people the system touches The AI-specific element frameworks require Likelihood How likely the risk is to occur One half of the risk score Severity How damaging the risk would be if it occurred The other half of the risk score Existing controls What already mitigates the risk Determines the residual risk Risk rating The resulting priority level Drives the order of treatment Treatment and owner The plan to address the risk and who owns it Turns assessment into accountable action The pattern across the fields is that the first four define and contextualize the risk, the middle three score it, and the last two convert it into action. A template missing the impact-on-individuals field fails the AI-specific requirement, one missing existing controls overstates risk by ignoring mitigation already in place, and one missing an owner produces analysis no one acts on. The fields that organizations most often omit, impact on individuals and a named owner, are precisely the ones that make the difference between an assessment that satisfies a framework and one that sits unused. The Scoring Approach The scoring approach in the template combines likelihood and severity into a risk rating, which is the mechanism that turns a list of risks into a priority order. Most templates use a simple matrix in which likelihood and severity are each rated on a scale, and their combination places each risk into a level such as low, medium, high, or critical. What matters is less the exact scale than that it is applied consistently, so that risks can be compared and the highest-rated ones addressed first. The AI-specific refinement is that severity should account for impact on individuals, not only impact on the organization. A risk with modest operational cost but serious potential to harm or unfairly affect people should score as significant, because the frameworks weigh that impact heavily. Building this into the severity judgment, rather than scoring purely on business impact, is what aligns the template’s prioritization with how AI governance expects risks to be ranked. The output is a rated, ordered set of risks that tells the organization where to act first. How the Template Aligns to ISO 42001 For organizations working to ISO 42001, an AI risk assessment is not optional but required: the standard’s planning clause requires both an AI risk assessment and an AI impact assessment, and a template that captures both satisfies that requirement directly. The risk fields support the risk assessment, the impact-on-individuals field supports the impact assessment, and the treatment and control fields feed the risk treatment and the Statement of Applicability that ISO 42001 expects. A template built around these fields therefore produces exactly the documented information an ISO 42001 audit looks for. The connection to the wider standard is covered in the guide to ISO 42001 requirements, which shows how the risk and impact assessments sit within the management system, and the specific design of the impact assessment is covered in the guide to designing the AI impact assessment for ISO 42001. The AIMS Manual provides the broader documented-information

EU AI Act Penalties: What Violations Cost and Who Pays

EU AI Act penalties are among the steepest in technology regulation, reaching, for the most serious violations, up to 7% of a company’s total worldwide annual turnover, a ceiling that exceeds even the headline fines under Europe’s data protection regime. That scale is deliberate, meant to make non-compliance a board-level financial risk rather than a cost of doing business. This guide explains the penalty tiers, who is liable across the AI value chain, and the compliance moves that keep an organization out of the highest-risk categories, so the exposure becomes something to manage rather than fear. The reason to understand the penalties before the deadlines arrive is that the compliance work they reward takes time to build, and an organization that waits until enforcement is imminent has the least room to reduce its exposure. The fines are the visible edge of the risk, but the practical response is a governance program that classifies AI use correctly and evidences responsible management, which is entirely achievable with lead time. The specific figures below are set by Article 99 of the Regulation, and because precision matters with statutory fines, this guide states them as the Regulation defines them. The EU AI Act Penalties Explained The EU AI Act structures its fines into tiers by the seriousness of the violation, and each tier is expressed as the higher of a fixed amount or a percentage of worldwide annual turnover. The table sets out the tiers as defined in Article 99 of the Regulation. Violation category Maximum penalty (higher of amount or turnover share) Prohibited AI practices Up to EUR 35 million or 7% of total worldwide annual turnover Non-compliance with other obligations (such as high-risk system requirements and transparency duties) Up to EUR 15 million or 3% of total worldwide annual turnover Supplying incorrect, incomplete, or misleading information to authorities Up to EUR 7.5 million or 1% of total worldwide annual turnover The structure tells a clear story: the law reserves its harshest penalties for the practices it prohibits outright, applies a substantial middle tier to failures in meeting the obligations placed on permitted but regulated systems, and sets a lower tier for information and cooperation failures. Because each ceiling is the higher of a fixed sum or a turnover percentage, the percentage bites hardest on large companies while the fixed amount sets a floor for smaller ones, which is why the exposure is genuinely material regardless of size. These figures are the statutory maximums set by Article 99 of the Regulation, defining the outer bound of what a violation can cost. Who Is Liable Across the AI Value Chain The EU AI Act does not place all liability on one party; it assigns obligations along the AI value chain, so who pays depends on the role a company plays. A provider, the party that develops an AI system or has it developed and places it on the market under its own name, carries the most extensive obligations, particularly for high-risk systems. A deployer, the party that uses an AI system in the course of its activities, carries its own distinct set of obligations around how the system is operated. Importers and distributors that bring AI systems into or move them through the market carry further responsibilities. The practical implication is that an organization must first understand which role or roles it occupies, because a company can be a provider of one system and a deployer of another, and its obligations, and therefore its exposure, follow accordingly. Misjudging the role is a common way organizations underestimate their liability, assuming that using a third-party AI tool carries no obligations when the deployer role in fact carries several. Mapping the roles accurately is the first step in understanding where the penalties could actually land. How the Fines Are Calculated Beyond the tier ceilings, the actual fine in any case is set with reference to the circumstances, so the maximums are a cap rather than a default. Authorities are directed to consider factors such as the nature and gravity of the violation, whether it was intentional or negligent, the steps taken to mitigate harm, and the degree of cooperation, which means an organization’s conduct materially affects where within the range a fine falls. A good-faith program that detects and remediates an issue is treated very differently from willful or concealed non-compliance. The Act also builds in proportionality for smaller organizations. For small and medium enterprises and startups, the fine is generally capped at the lower of the applicable percentage or the fixed amount, rather than the higher, which prevents a single penalty from being existential for a small company in the way it could be for a large one. The principle is that the regime scales with the organization, so the exposure a company faces is a function of its size, its role, and above all its conduct. What Exposure Actually Looks Like Focusing only on the headline fine understates the exposure, because the financial penalty is one part of what non-compliance costs. Enforcement can also bring orders to withdraw or recall an AI system from the market, which for a company whose product is the AI system is a direct hit to revenue rather than a one-time fine. There is reputational damage that follows public enforcement, the operational cost of remediating under a deadline set by a regulator, and the commercial friction of customers and partners reassessing the relationship once a violation is known. Seen this way, the real exposure is to the business, not just the balance sheet, which is what makes the penalties a board-level concern rather than a compliance line item. It also reframes the value of compliance: the return is not only avoiding a fine but preserving market access, reputation, and customer trust. An organization that treats EU AI Act penalties as the whole risk misses the larger picture, while one that manages the underlying compliance protects against all of it at once. The Compliance Moves That Cap Exposure The moves