Skip to main content
    Kriv AI

    AI governance

    AI accountability: who answers when an AI system gets it wrong

    A plain guide for leaders who must say who is responsible for an AI system. The wording comes from NIST's AI Risk Management Framework, read on 11 November 2026. The framework is voluntary and NIST says it is being updated, so check the current version.

    AI accountability means that specific people and teams answer for what an AI system does, and can be asked about it. NIST's AI Risk Management Framework says trustworthy AI depends on accountability and that accountability presupposes transparency, meaning information about the system and its outputs is available. It also says the roles of the people involved should be considered when seeking accountability for outcomes. In practice, a team writes down who owns each AI system, who can review its outputs, and who decides when it should stop. In Kriv's view, a vendor or a model should not hold that role for you.

    plain answer

    What is AI accountability?

    AI accountability is the arrangement that lets an organization say who answers for an AI system's outputs and what happens when they are wrong. NIST's AI Risk Management Framework states that trustworthy AI depends upon accountability, and that accountability presupposes transparency. NIST defines transparency as the extent to which information about an AI system and its outputs is available.

    NIST also says transparency is often necessary for actionable redress when outputs are incorrect or lead to negative impacts. Put simply, if nobody can see what the system did, nobody can be held to account for it.

    Sources: NIST AIRC, AI Risks and Trustworthiness, Section 3.4, accountability and transparency (https://airc.nist.gov/airmf-resources/airmf/3-sec-characteristics/), read 11 November 2026.

    plain answer

    Who is accountable for an AI system?

    Who is accountable for an AI system is a question NIST answers by role, not by name. Its framework says the role of AI actors should be considered when seeking accountability for the outcomes of AI systems, and that it is the joint responsibility of all AI actors to decide whether AI is an appropriate or necessary tool. In its Govern function, NIST says executive leadership takes responsibility for decisions about risks associated with AI system development and deployment.

    The framework does not name a single job title. In plain terms, the people who build, buy, deploy and use a system each hold part of the answer, and leadership holds the decision about the risk. How your organization divides those parts is for your own leaders and counsel to decide.

    Sources: NIST AIRC, Section 3 introduction and Section 3.4 (https://airc.nist.gov/airmf-resources/airmf/3-sec-characteristics/) and Section 5, AI RMF Core, Govern 2.3 (https://airc.nist.gov/airmf-resources/airmf/5-sec-core/), read 11 November 2026. The last sentence is Kriv's plain reading.

    plain answer

    How does AI accountability work in an organization?

    AI accountability works through written roles and clear lines of communication. NIST's Govern 2 category says accountability structures should be in place so that the appropriate teams and individuals are empowered, responsible and trained for mapping, measuring and managing AI risks. Govern 2.1 adds that roles, responsibilities and lines of communication are documented and clear to individuals and teams throughout the organization.

    That wording points to documents, not intentions: a named owner, a described role, and people who know the path for raising a problem.

    Sources: NIST AIRC, Section 5, AI RMF Core, Govern 2 and Govern 2.1 (https://airc.nist.gov/airmf-resources/airmf/5-sec-core/), read 11 November 2026. The last sentence is Kriv's reading.

    how to

    How can a team set up AI accountability?

    A team can set up AI accountability by answering a short list of questions for each AI system and writing the answers down. These are Kriv's suggestions, drawn from NIST's wording, not NIST's own procedure.

    Who owns this system, by name and role? Who reviews its outputs, and how often? Who can see what it did, and where is that recorded? Who decides when it is paused or switched off? Who do users tell when it is wrong, and who answers them? Which executive accepts the remaining risk?

    Keep the answers in one register, review them on a set date, and update them when the system or its use changes. NIST says developers and deployers should consider adjusting transparency and accountability practices in proportion when the consequences are severe, so a system with serious consequences deserves more than a system with minor ones.

    Sources: NIST AIRC, Section 3.4 (https://airc.nist.gov/airmf-resources/airmf/3-sec-characteristics/), read 11 November 2026. The questions are Kriv's suggestions.

    plain answer

    How is AI accountability different from AI governance?

    AI accountability is one part of AI governance. Governance is the wider set of policies, roles, controls and records, and accountability is the part that says who answers. NIST lists accountability structures as a category inside its Govern function, which supports that reading.

    Our AI governance framework page covers where to start from a blank page, and our AI governance roles page covers who fills these roles.

    Sources: NIST AIRC, Section 5 (https://airc.nist.gov/airmf-resources/airmf/5-sec-core/), read 11 November 2026; Kriv AI live pages.

    kriv role

    Where Kriv helps

    Kriv AI's AI governance consulting helps an organization write down owners, review steps and escalation paths for its AI systems, and builds the register and monitoring that keep them current. Kriv AI sends a senior practitioner who both scopes and delivers.

    Kriv does not certify compliance and makes no claim of formal alignment with the NIST framework. This page is informed by NIST's published wording. Your own leaders and counsel decide who answers for each system.

    limits

    What this page could not confirm

    We read two NIST pages, Section 3.4 and Section 5 of AI RMF 1.0, and not the full framework document, whose PDF we could not extract. We did not read other bodies of guidance, laws or court decisions on liability, and nothing here says who is legally liable for an AI system. We give no statistics. NIST says AI RMF 1.0 is being updated, so the wording may change. We have no named client case study on this topic. For anything not cited above, check NIST's current pages.

    Straight answers

    Frequently asked questions about AI accountability: who answers when an AI system gets it wrong

    What does NIST say about AI accountability?

    NIST says trustworthy AI depends upon accountability and that accountability presupposes transparency (NIST AI RMF 1.0, Section 3.4, read 11 November 2026).

    Who is responsible for AI decisions?

    NIST says the role of AI actors should be considered when seeking accountability for outcomes, and that executive leadership takes responsibility for decisions about AI risks (Sections 3.4 and 5, read 11 November 2026).

    Can a vendor be accountable for our AI system?

    NIST's Govern 6 category covers policies for AI risks arising from third-party software and data (AI RMF Core, Section 5, read 11 November 2026). Who answers legally is a question for your counsel.

    Is the NIST framework mandatory?

    NIST describes the AI Risk Management Framework as intended for voluntary use (NIST AI RMF page, read 11 November 2026).

    Can Kriv take on accountability for our AI?

    No. Kriv AI helps write and run the governance around an AI system, but your organization's leaders remain accountable. Kriv does not certify compliance or promise outcomes.

    Talk to the team that would do the work

    Bring your requirements to a working session with the person who'll actually deliver.

    Book a Discovery Call