owasp-llm-checklist.yaml

urn: urn:intuitem:risk:library:owasp-llm-checklist
locale: en
ref_id: owasp-llm-checklist
name: LLM AI Cybersecurity & Governance Checklist
description: This checklist is intended to help technology and business leaders quickly
  understand the risks and bene ts of using LLM, allowing them to focus on developing
  a comprehensive list of critical areas and tasks needed to defend and protect the
  organization as they develop a Large Language Model strategy.
copyright: OWASP - Creative Commons Attribution-ShareAlike 4.0
version: 1
provider: OWASP
packager: intuitem
objects:
  framework:
    urn: urn:intuitem:risk:framework:owasp-llm-checklist
    ref_id: owasp-llm-checklist
    name: LLM AI Cybersecurity & Governance Checklist
    description: This checklist is intended to help technology and business leaders
      quickly understand the risks and bene ts of using LLM, allowing them to focus
      on developing a comprehensive list of critical areas and tasks needed to defend
      and protect the organization as they develop a Large Language Model strategy.
    requirement_nodes:
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:1
      assessable: false
      depth: 1
      ref_id: '1'
      name: Adversarial Risk
      description: Adversarial Risk includes competitors and attackers.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:1.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:1
      ref_id: '1.1'
      description: 'Scrutinize how competitors are investing in artificial intelligence.
        Although there are risks in AI

        adoption, there are also business benefits that may impact future market positions.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:1.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:1
      ref_id: '1.2'
      description: 'Investigate the impact of current controls, such as password resets,
        which use voice

        recognition which may no longer provide the appropriate defensive security
        from new GenAI

        enhanced attacks.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:1.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:1
      ref_id: '1.3'
      description: 'Update the Incident Response Plan and playbooks for GenAI enhanced
        attacks and AIML

        specific incidents.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      assessable: false
      depth: 1
      ref_id: '2'
      name: Threat Modeling
      description: 'Threat modeling is highly recommended to identify threats and
        examine processes and security

        defenses. Threat modeling is a set of systematic, repeatable processes that
        enable making

        reasonable security decisions for applications, software, and systems. Threat
        modeling for GenAI

        accelerated attacks and before deploying LLMs is the most cost effective way
        to Identify and mitigate

        risks, protect data, protect privacy, and ensure a secure, compliant integration
        within the business.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      ref_id: '2.1'
      description: 'How will attackers accelerate exploit attacks against the organization,
        employees, executives,

        or users? Organizations should anticipate "hyper-personalized" attacks at
        scale using

        Generative AI. LLM-assisted Spear Phishing attacks are now exponentially more
        effective,

        targeted, and weaponized for an attack.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      ref_id: '2.2'
      description: How could GenAI be used for attacks on the business's customers
        or clients through spoofing or GenAI generated content?
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      ref_id: '2.3'
      description: Can the business detect and neutralize harmful or malicious inputs
        or queries to LLM solutions?
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2.4
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      ref_id: '2.4'
      description: Can the business safeguard connections with existing systems and
        databases with secure integrations at all LLM trust boundaries?
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2.5
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      ref_id: '2.5'
      description: Does the business have insider threat mitigation to prevent misuse
        by authorized users?
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2.6
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      ref_id: '2.6'
      description: Can the business prevent unauthorized access to proprietary models
        or data to protect Intellectual Property?
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2.7
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:2
      ref_id: '2.7'
      description: Can the business prevent the generation of harmful or inappropriate
        content with automated content filtering?
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3
      assessable: false
      depth: 1
      ref_id: '3'
      name: AI Asset Inventory
      description: An AI asset inventory should apply to both internally developed
        and external or third-party solutions.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3
      ref_id: '3.1'
      description: Catalog existing AI services, tools, and owners. Designate a tag
        in asset management for specific inventory.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3
      ref_id: '3.2'
      description: Include AI components in the Software Bill of Material (SBOM),
        a comprehensive list of all the software components, dependencies, and metadata
        associated with applications.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3
      ref_id: '3.3'
      description: Catalog AI data sources and the sensitivity of the data (protected,
        confidential, public)
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3.4
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3
      ref_id: '3.4'
      description: Establish if pen testing or red teaming of deployed AI solutions
        is required to determine the current attack surface risk.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3.5
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3
      ref_id: '3.5'
      description: Create an AI solution onboarding process.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3.6
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:3
      ref_id: '3.6'
      description: Ensure skilled IT admin staff is available either internally or
        externally, following SBoM requirements.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4
      assessable: false
      depth: 1
      ref_id: '4'
      name: AI Security and Privacy Training
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4
      ref_id: '4.1'
      description: Actively engage with employees to understand and address concerns
        with planned LLM initiatives.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4
      ref_id: '4.2'
      description: Establish a culture of open, and transparent communication on the
        organization's use of predictive or generative AI within the organization
        process, systems, employee management and support, and customer engagements
        and how its use is governed, managed, and risks addressed.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4
      ref_id: '4.3'
      description: Train all users on ethics, responsibility, and legal issues such
        as warranty, license, and copyright.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4.4
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4
      ref_id: '4.4'
      description: Update security awareness training to include GenAI related threats.
        Voice cloning and image cloning, as well as in anticipation of increased spear
        phishing attacks
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4.5
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:4
      ref_id: '4.5'
      description: Any adopted GenAI solutions should include training for both DevOps
        and cybersecurity for the deployment pipeline to ensure AI safety and security
        assurances.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:5
      assessable: true
      depth: 1
      ref_id: '5'
      name: Establish Business Cases
      description: Solid business cases are essential to determining the business
        value of any proposed AI solution, balancing risk and bene fits, and evaluating
        and testing return on investment. There are an enormous number of potential
        use cases; a few examples are provided.
      annotation: '* Enhance customer experience

        * Better knowledge management

        * Better operational ef ficiency

        * Enhanced innovation

        * Document creation, translation, summarization, and analysis

        * Market Research and Competitor Analysis'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6
      assessable: false
      depth: 1
      ref_id: '6'
      name: Governance
      description: Corporate governance in LLM is needed to provide organizations
        with transparency and accountability. Identifying AI platform or process owners
        who are potentially familiar with the technology or the selected use cases
        for the business is not only advised but also necessary to ensure adequate
        reaction speed that prevents collateral damages to well established enterprise
        digital processes.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6
      ref_id: '6.1'
      description: Establish the organization's AI RACI chart (who is responsible,
        who is accountable, who should be consulted, and who should be informed)
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6
      ref_id: '6.2'
      description: Document and assign AI risk, risk assessments, and governance responsibility
        within the organization.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6
      ref_id: '6.3'
      description: Establish data management policies, including technical enforcement,
        regarding data classification and usage limitations. Models should only leverage
        data classified for the minimum access level of any user of the system. For
        example, update the data protection policy to emphasize not to input protected
        or confidential data into non business-managed tools.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6.4
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6
      ref_id: '6.4'
      description: Create an AI Policy supported by established policy (e.g., standard
        of good conduct, data protection, software use)
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6.5
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6
      ref_id: '6.5'
      description: Publish an acceptable use matrix for various generative AI tools
        for employees to use.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6.6
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:6
      ref_id: '6.6'
      description: Document the sources and management of any data that the organization
        uses from the generative LLM models.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      assessable: false
      depth: 1
      ref_id: '7'
      name: Legal
      description: Many of the legal implications of AI are unde fined and potentially
        very costly. An IT, security, and legal partnership is critical to identifying
        gaps and addressing obscure decisions.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.1'
      description: Confirm product warranties are clear in the product development
        stream to assign who is responsible for product warranties with AI.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.2'
      description: Review and update existing terms and conditions for any GenAI considerations.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.3'
      description: Review AI EULA agreements. End-user license agreements for GenAI
        platforms are very different in how they handle user prompts, output rights
        and ownership, data privacy, compliance, liability, privacy, and limits on
        how output can be used.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.4
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.4'
      description: Organizations EULA for customers, Modify end-user agreements to
        prevent the organization from incurring liabilities related to plagiarism,
        bias propagation, or intellectual property infringement through AI-generated
        content.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.5
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.5'
      description: Review existing AI-assisted tools used for code development. A
        chatbot's ability to write code can threaten a company's ownership rights
        to its product if a chatbot is used to generate code for the product. For
        example, it could call into question the status and protection of the generated
        content and who holds the right to use the generated content.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.6
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.6'
      description: Review any risks to intellectual property. Intellectual property
        generated by a chatbot could be in jeopardy if improperly obtained data was
        used during the generative process, which is subject to copyright, trademark,
        or patent protection. If AI products use infringing material, it creates a
        risk for the outputs of the AI, which may result in intellectual property
        infringement.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.7
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.7'
      description: Review any contracts with indemni fication provisions. Indemni
        fication clauses try to put the responsibility for an event that leads to
        liability on the person who was more at fault for it or who had the best chance
        of stopping it. Establish guardrails to determine whether the provider of
        the AI or its user caused the event, giving rise to liability.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.8
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.8'
      description: Review liability for potential injury and property damage caused
        by AI systems.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.9
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.9'
      description: Review insurance coverage. Traditional (D&O) liability and commercial
        general liability insurance policies are likely insufficient to fully protect
        AI use.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.10
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.10'
      description: Identify any copyright issues. Human authorship is required for
        copyright. An organization may also be liable for plagiarism, propagation
        of bias, or intellectual property infringement if LLM tools are misused.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.11
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.11'
      description: Ensure agreements are in place for contractors and appropriate
        use of AI for any development or provided services.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.12
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.12'
      description: Restrict or prohibit the use of generative AI tools for employees
        or contractors where enforceable rights may be an issue or where there are
        IP infringement concerns.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.13
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.13'
      description: Assess and AI solutions used for employee management or hiring
        could result in disparate treatment claims or disparate impact claims.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7.14
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:7
      ref_id: '7.14'
      description: Make sure the AI solutions do not collect or share sensitive information
        without proper consent or authorization.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      assessable: false
      depth: 1
      ref_id: '8'
      name: Regulatory
      description: "The EU AI Act is anticipated to be the first comprehensive AI\
        \ law but will apply in  2025 at the earliest. The EU\u2019s General Data\
        \ Protection Regulation (GDPR) does not speci cally address AI but includes\
        \ rules for data collection, data security, fairness and transparency, accuracy\
        \ and reliability, and accountability, which can impact GenAI use. In the\
        \ United States, AI regulation is included within broader consumer privacy\
        \ laws. Ten US states have passed laws or have laws that will go into effect\
        \ by the end of  2023. Canada has so far only published a Voluntary Code of\
        \ Conduct on the Responsible Development and Management of Advanced Generative\
        \ AI Systems, however, the Artificial Intelligence and Data Act (AIDA) will\
        \ have stronger requirements. Federal organizations such as the US Equal Employment\
        \ Opportunity Commission (EEOC), the Consumer Financial Protection Bureau\
        \ (CFPB), the Federal Trade Commission (FTC), and the US Department of Justice's\
        \ Civil Rights Division (DOJ) are closely monitoring hiring fairness."
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.1'
      description: ' Determine Country, State, or other Government specific AI compliance
        requirements.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.2'
      description: ' Determine compliance requirements for restricting electronic
        monitoring of employees and employment-related automated decision systems
        (Vermont, California, Maryland, New York, New Jersey)'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.3'
      description: ' Determine compliance requirements for consent for facial recognition
        and the AI video analysis required (Illinois, Maryland, Washington, Vermont)'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.4
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.4'
      description: ' Confirm the vendor''s compliance with applicable AI laws and
        best practices.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.5
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.5'
      description: ' Review any AI tools in use or being considered for employee hiring
        or management.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.6
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.6'
      description: ' Ask and document any products using AI during the hiring process.
        Ask how the model was trained, and how it is monitored, and track any corrections
        made to avoid discrimination and bias.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.7
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.7'
      description: ' Ask and document what accommodation options are included.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.8
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.8'
      description: ' Ask how the vendor or tool stores and deletes data and regulates
        the use of facial recognition and video analysis tools during pre-employment.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.9
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.9'
      description: ' Ask and document whether the vendor collects confidential data.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8.10
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:8
      ref_id: '8.10'
      description: ' Review other organization-specific regulatory requirements with
        AI that may raise compliance issues. The Employee Retirement Income Security
        Act of 1974, for instance, has fiduciary duty requirements for retirement
        plans that a chatbot might not be able to meet.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      assessable: false
      depth: 1
      ref_id: '9'
      name: Using or Implementing Large Language Model Solutions
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.1'
      description: Threat Model LLM components and architecture trust boundaries.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.2'
      description: Data Security, verify how data is classi fied and protected based
        on sensitivity, including personal and proprietary business data. (How are
        user permissions managed, and what safeguards are in place?)
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.3
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.3'
      description: Access Control, implement least privilege access controls and implement
        defense-in-depth measures
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.4
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.4'
      description: Training Pipeline Security, require rigorous control around training
        data governance, pipelines, models, and algorithms.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.5
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.5'
      description: Input and Output Security, evaluate input validation methods, as
        well as how outputs are filtered, sanitized, and approved.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.6
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.6'
      description: Monitoring and Response, map workflows, monitoring, and responses
        to understand automation, logging, and auditing. Con firm audit records are
        secure.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.7
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.7'
      description: Include application testing, source code review, vulnerability
        assessments, and red teaming in the production release process.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.8
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.8'
      description: Check for existing vulnerabilities in the LLM model or supply chain.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.9
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.9'
      description: Look into the effects of threats and attacks on LLM solutions,
        such as prompt injection, the release of sensitive information, and process
        manipulation.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.10
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.10'
      description: Investigate the impact of attacks and threats to LLM models, including
        model poisoning, improper data handling, supply chain attacks, and model theft.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.11
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.11'
      description: Supply Chain Security, request third-party audits, penetration
        testing, and code reviews for third-party providers. (both initially and on
        an ongoing basis)
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.12
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.12'
      description: Infrastructure Security, ask how often a vendor performs resilience
        testing? What are their SLAs in terms of availability, scalability, and performance?
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.13
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.13'
      description: Update incident response playbooks and include an LLM incident
        in tabletop exercises.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9.14
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:9
      ref_id: '9.14'
      description: Identify or expand metrics to benchmark generative cybersecurity
        AI against other approaches to measure expected productivity improvements.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:10
      assessable: false
      depth: 1
      ref_id: '10'
      name: Testing, Evaluation, Veri fication, and Validation (TEVV)
      description: NIST AI Framework recommends a continuous TEVV process throughout
        the AI lifecycle which includes the AI system operators, domain experts, AI
        designers, users, product developers, evaluators, and auditors. TEVV includes
        a range of tasks such as system validation, integration, testing, recalibration,
        and ongoing monitoring for periodic updates to navigate the risks and changes
        of the AI system.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:10.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:10
      ref_id: '10.1'
      description: Establish continuous testing, evaluation, verification, and validation
        throughout the AI model lifecycle.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:10.2
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:10
      ref_id: '10.2'
      description: Provide regular executive metrics and updates on AI Model functionality,
        security, reliability, and robustness.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      assessable: false
      depth: 1
      ref_id: '11'
      name: Model Cards and Risk Cards
      description: Model cards and risk cards are foundational elements for increasing
        the transparency, accountability, and ethical deployment of Large Language
        Models (LLMs). Model cards help users understand and trust AI systems by providing
        standardized documentation on their design, capabilities, and constraints,
        leading them to make educated and safe applications. Risk cards supplement
        this by openly addressing potential negative consequences, such as biases,
        privacy problems, and security vulnerabilities, which encourages a proactive
        approach to harm prevention. These documents are critical for developers,
        users, regulators, and ethicists equally since they establish a collaborative
        atmosphere in which AI's social implications are carefully addressed and handled.
        These cards, developed and maintained by the organizations that created the
        models, play an important role in ensuring that AI technologies ful fill ethical
        standards and legal requirements, allowing for responsible research and deployment
        in the AI ecosystem.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.1
      assessable: false
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.1'
      description: 'Model details : Basic information about the model, i.e., name,
        version, and type ( neural network, decision tree, etc.), and the intended
        use case.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.2
      assessable: false
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.2'
      description: 'Model architecture : Includes a description of the structure of
        the model, such as the number and type of layers, activation functions, and
        other key architectural choices.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.3
      assessable: false
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.3'
      description: 'Training data and methodology : Information about the data used
        to train the model, such as the size of the dataset, the data sources, and
        any preprocessing or data augmentation techniques used. It also includes details
        about the training methodology, such as the optimizer used, the loss function,
        and any hyperparameters that were tuned.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.4
      assessable: false
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.4'
      description: 'Performance metrics : Information about the model''s performance
        on various metrics, such as accuracy, precision, recall, and F /one.pnum score.
        It may also include information about how the model performs on different
        subsets of the data.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.5
      assessable: false
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.5'
      description: 'Potential biases and limitations : Lists potential biases or limitations
        of the model, such as imbalanced training data, over fitting, or biases in
        the model''s predictions. It may also include information about the model''s
        limitations, such as its ability to generalize to new data or its suitability
        for certain use cases.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.6
      assessable: false
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.6'
      description: 'Responsible AI considerations : Any ethical or responsible AI
        considerations related to the model, such as privacy concerns, fairness, and
        transparency, or potential societal impacts of the model''s use. It may also
        include recommendations for further testing, validation, or monitoring of
        the model.'
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.7
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.7'
      description: Review a model's model card
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.8
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.8'
      description: Review risk card if available
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11.9
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:11
      ref_id: '11.9'
      description: Establish a process to track and maintain model cards for any deployed
        model including models used through a third party.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:12
      assessable: true
      depth: 1
      ref_id: '12'
      name: 'RAG: Large Language Model Optimization'
      description: Retrieval-Augmented Generation RAG has evolved as a more effective
        way of optimizing and augmenting the capabilities of large language models
        by retrieving pertinent data from up to date available knowledge sources.
        RAG can be customized for speci fic domains, optimizing the retrieval of domain-speci
        fic information and tailoring the generation process to the nuances of specialized
        fields. RAG is seen as a more ef ficient and transparent method for LLM optimization,
        particularly for problems where labeled data is limited or expensive to collect.
        One of the primary advantages of RAG is its support for continuous learning
        since new information can be continually updated at the retrieval stage.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:13
      assessable: false
      depth: 1
      ref_id: '13'
      name: AI Red Teaming
      description: AI Red Teaming is an adversarial attack test simulation of the
        AI System to validate there aren't any existing vulnerabilities which can
        be exploited by an attacker. It is a recommended practice by many regulatory
        and AI governing bodies including the Biden administration. Red-teaming alone
        is not a comprehensive solution to validate all real-world harms associated
        with AI systems and should be included with other forms of testing, evaluation,
        veri fication, and validation such as algorithmic impact assessments and external
        audits.
    - urn: urn:intuitem:risk:req_node:owasp-llm-checklist:13.1
      assessable: true
      depth: 2
      parent_urn: urn:intuitem:risk:req_node:owasp-llm-checklist:13
      ref_id: '13.1'
      description: Incorporate Red Team testing as a standard practice for AI Models
        and applications.