Risk and compliance tab in AI Control Tower
Summarize
Summary of Risk and compliance tab in AI Control Tower
The Risk and compliance tab in AI Control Tower offers ServiceNow customers a centralized dashboard to monitor and manage the risk classification and compliance posture of AI assets. These assets include AI systems, AI models, and datasets essential for AI development, deployment, and operation. Understanding and managing these assets within the tab helps ensure effectiveness, reliability, and adherence to regulatory and ethical standards.
Show less
Key Features
- Compliance Overview: Displays risk classifications of AI assets (High, Medium, Low, Unacceptable) via visual donut charts. Shows compliance scores for authority documents and policies based on controls implemented.
- Compliance Frameworks: Supports frameworks such as NIST AI Risk Management Framework and the EU Artificial Intelligence Act. Users can view compliance data by selecting authority documents or policies and filter for overall compliance scores, compliant/non-compliant counts, and related issues.
- Risk Overview: Tracks risk posture of AI assets with filters by asset type. Aggregated risk scores for AI systems are visualized as High or Low risk via donut charts.
- Risk Heatmap: Visualizes identified risks segmented by risk and control effectiveness or impact and likelihood. Allows filtering by residual or inherent risk and by multiple Risk Assessment Methodologies (RAMs) if applicable.
- Regulatory Landscape Overview (requires additional app): Displays regulatory alerts, change tasks, and assessments by workflow state or lifecycle phase, supporting proactive regulatory change management.
- Configurable Authority Documents and Policies: Customers can select which authority documents and policies to display on the dashboard’s home page, tailoring the view to their compliance needs.
Key Outcomes
By utilizing the Risk and compliance tab, ServiceNow customers can gain comprehensive visibility into AI asset risks and compliance status mapped to recognized regulatory frameworks. This enables informed decision-making, prioritization of risk mitigation efforts, and streamlined compliance management. The solution supports proactive monitoring of regulatory changes and related impact assessments, helping organizations maintain alignment with applicable laws and standards while managing AI governance effectively.
The Risk and compliance tab on the AI Control Tower displays the risk classification of an AI asset inventory and the compliance posture for the selected authority documents and policies.
AI assets refer to the various components and resources that are essential for the development, deployment, and operation of artificial intelligence systems. These assets can include:
- AI systems The complete software or hardware infrastructure that runs AI algorithms and processes. This can include machine learning platforms, natural language processing systems, and other AI-driven applications.
- AI models: The mathematical and computational models that are trained on data to perform specific tasks. These models can range from simple linear regression models to complex deep learning neural networks.
- Datasets: The collections of data used to train, validate, and test AI models.
Understanding and managing these AI assets is crucial for ensuring that AI systems are effective, reliable, and compliant with regulatory and ethical standards as defined by your organization.
The Risk & compliance dashboard has the following sections. You can drill-down into the data on each widget in any section.
Compliance overview
- Regulatory risk classification
- This section displays the risk classifications of AI systems, AI models, Datasets using donut charts. The risks are qualitatively classified as High, Low, Medium, and Unacceptable. These classifications are based on the risk assessments of the AI assets.
- Compliance by authority documents and policies
- The section shows compliance based on controls implemented. By default, the compliance scores are displayed for the following frameworks that are provided in the library:
- NIST AI Risk Management Framework: This framework displays the four key associated citations, namely map, measure, manage, and govern. Each citation's compliance score is displayed based on its control attestations.
- EU Artificial Intelligence Act: This framework has multiple chapters that are displayed as citations and child citations. Each citation is mapped to a control objective to provide you with a compliance percentage score.
You can choose to view compliance data by selecting one of two options: Authority Documents or Policies. Additionally, you can view the overall compliance score percentage, along with the number of compliant and non-compliant authority documents and policies, by using the drop-down filter to select specific authority documents or policies. You can also see all the issues that require immediate attention and AI cases related to each authority document or policy.
The authority documents are provided solely for informational and guidance purposes to assist with the initial setup of AI Risk and Compliance frameworks. It does not constitute legal advice or assurance of regulatory compliance. You are solely responsible for ensuring that all use of the content complies with applicable laws, regulations, directives, and industry standards in their jurisdictions.
Note:You can configure which authority documents and policies you want to display on the home page. For more information, see Set up properties for compliance posture.
Risk overview
This section monitors and tracks the risk posture of the AI assets in your organization. Using the AI asset filter, you can filter risk posture insights by the type of AI asset inventory.
- AI systems by aggregated risk score
- This section displays the classifications of AI systems by aggregated risk score using donut chart. The risk scores are qualitatively classified as High and Low.
- Risk heatmap
- The Risk heatmap widget displays the visualization of all identified risks within the AI assets. By default, residual risk filter is applied, but you can filter it based on inherent risk level. The heatmap is segmented, and the segmentation changes based on the filter. The activities fall under the respective combination of risk and control effectiveness, or impact and likelihood. The combination is based on the selected risk classification filter. You can filter the risk heatmap by Risk Assessment Methodology (RAM), if you have more than one risk RAMs published.
Regulatory landscape overview
You need to install GRC: Regulatory Change Management application to see this section. For more information, see Installing Regulatory Change Management.
- Overview
-
- Alerts
This section displays the distribution of regulatory alerts by workflow state or lifecycle phase using a donut chart.
- Change tasks
This section displays the distribution of regulatory change tasks by workflow state or lifecycle phase using a donut chart.
- Alerts
- Assessments
- This section displays the regulatory impact assessments and risk assessments linked to regulatory changes using a donut chart. By default, regulatory assessments filter is applied, but you can change it to risk assessments.
The following image shows the Risk & compliance dashboard.
For more information, see AI Risk and Compliance documentation.