AI Security ダッシュボード(プレビュー)には、AI セキュリティ ポスチャーの概要が表示されます。
ダッシュボードには、次のセクションが表示されます。
Riskiest AI Issues: AI インベントリの最もリスクの高い問題を、攻撃の発生可能性スコアの高い順に表示します。問題をクリックすると、詳細が表示されます。
AI インベントリ: 生成 AI アクティビティ、使用されているモデル、データセットを含むプロジェクトの概要を確認します。ノードをクリックすると詳細が表示されます。Sensitive Data Protection が有効になっている場合、ダッシュボードにはデータセットに機密データが含まれているかどうかが表示されます。
検出結果: AI セキュリティ ポリシーとデータ セキュリティ ポリシーによって生成された検出結果を評価して管理します。
Vertex AI データセット内の機密データ: Vertex AI データセット内の機密データの検出結果を評価して管理します。
Model Armor の検出結果: Model Armor によってスキャンされたプロンプトとレスポンスのグラフを表示します。プロンプト インジェクションや機密データの検出などの検出された問題が表示されます。
[[["わかりやすい","easyToUnderstand","thumb-up"],["問題の解決に役立った","solvedMyProblem","thumb-up"],["その他","otherUp","thumb-up"]],[["わかりにくい","hardToUnderstand","thumb-down"],["情報またはサンプルコードが不正確","incorrectInformationOrSampleCode","thumb-down"],["必要な情報 / サンプルがない","missingTheInformationSamplesINeed","thumb-down"],["翻訳に関する問題","translationIssue","thumb-down"],["その他","otherDown","thumb-down"]],["最終更新日 2025-09-04 UTC。"],[],[],null,["| Enterprise [service tier](/security-command-center/docs/service-tiers) (requires [organization-level activation](/security-command-center/docs/activate-scc-overview#overview_of_organization-level_activation))\n\nThe **Risk** section in the Google Cloud console helps you to manage the highest\nprofile risks in your cloud environment.\n\nThe **Overview** page serves as your first contact security dashboard,\nhighlighting the highest priority risks in your cloud environments. You can view\nmultiple domains in **Overview** by selecting one of the following views:\n\n- [All risk](#all-risk)\n- [Vulnerabilities](#vulnerabilities)\n- [Code](#code)\n- [AI security](#ai-protection) ([Preview](/products#product-launch-stages))\n\nAll risk dashboard\n\nThe **All risk** dashboard surfaces the following high-priority security risks\nacross your cloud environments:\n\n- **Riskiest issues** , which shows at-a-glance information for your top [issues](/security-command-center/docs/issues-overview), including simplified attack paths or evidence diagrams.\n- **Recent Google Cloud threats** that are active in your cloud environments.\n- **High-impact, exploitable vulnerabilities**, prioritized by the resources impacted by them.\n\nThe dashboard also displays your progress in **Compliance**, like NIST, HIPAA,\nPCI-DSS, and CIS, reported as a percentage of how many controls are passing.\n\nIn most cases you can interact with individual high priority risks for a brief\nsummary, continue on to a more detailed view of each risk, or view all risks of\na specific type.\n\nVulnerabilities dashboard\n\nThe **Vulnerabilities** dashboard gives insights into virtual machines and\ncontainers with exploitable vulnerabilities across your cloud environments. The\ndashboard displays the following information:\n\n- **Top common vulnerabilities and exploits** . Displays a clickable quadrant\n heatmap to help you filter vulnerabilities by\n [exploitability](/security-command-center/docs/reference/rest/v2/organizations.sources.findings#exploitationactivity)\n and impact [(risk\n rating)](/security-command-center/docs/reference/rest/v2/organizations.sources.findings#riskrating).\n The number of unique resources affected, and the findings related to those\n resources are shown in a table after the heatmap. Each unique resource might\n have more than one finding.\n\n To reset the heatmap, click a heatmap cell again.\n- **Most common critical exploitable vulnerabilities**. A list of highly\n exploitable vulnerabilities found in your cloud environments, prioritized by\n the total number of unique resources impacted by them.\n\n Expand a CVE section to view its descriptions, which are findings related to\n the CVE and the resources they affect. Because different findings can affect\n the same resource, the sum of all resource counts in the expanded\n description might be greater than the unique resource count in the heading\n row.\n- **Containers with exploitable vulnerabilities** . A list of containers with\n exploitable vulnerabilities, where the vulnerability [exploitation\n activity](/security-command-center/docs/reference/rest/v2/organizations.sources.findings#exploitationactivity)\n rating is `available`, `confirmed`, or `wide` and the [risk rating](/security-command-center/docs/reference/rest/v2/organizations.sources.findings#Finding.RiskRating)\n is `critical`, based on the assessment of Google Threat Intelligence. The\n list is ordered by [attack exposure\n score](/security-command-center/docs/attack-exposure-learn), then by largest number of\n impacted resources.\n\n- **Latest compute vulnerabilities with known exploits** . A list of\n Compute Engine virtual machine instances that have exploitable\n vulnerabilities with findings that belong to the `OS_VULNERABILITY` or\n `SOFTWARE_VULNERABILITY` category.\n\n From here you can check the following:\n - The attack exposure score of the exploit. Click the score to see the attack paths to your exposed [high-value\n resources](/security-command-center/docs/attack-exposure-learn#high-value_resources).\n - How many [configured high-value resources](/security-command-center/docs/attack-exposure-learn#resource-value-configurations) have been exposed due to the vulnerability, that have a [priority](/security-command-center/docs/attack-exposure-learn#priority-values-manual) of `HIGH`, `MEDIUM`, or `LOW`.\n - The **Exploit release date**, which is when the vulnerability was announced.\n - The **First available date**, which is when an exploit was first observed.\n - The [level of exploitability](/security-command-center/docs/reference/rest/v2/organizations.sources.findings#exploitationactivity) of the vulnerability.\n\nCode dashboard\n\nThe **Code** dashboard shows code vulnerabilities in your cloud environments\nfound by Snyk. To use it, you need to [set up the Snyk integration](/security-command-center/docs/how-to-configure-snyk-for-scc).\n\nYou can also [view code vulnerabilities in the Google Cloud console](/security-command-center/docs/code-security-findings-snyk).\n\nAI Security dashboard\n\nThe **AI Security** dashboard ([Preview](/products#product-launch-stages)) provides a\nhigh-level view of your AI security posture.\n\nThe dashboard displays the following sections:\n\n- **Riskiest AI Issues** : View top risks in your AI inventory, prioritized by the highest [attack exposure scores](/security-command-center/docs/attack-exposure-learn). Click any issue to see details.\n- **AI Inventory**: Get an overview of projects with generative AI activity, models used, and datasets. Click nodes for details. If Sensitive Data Protection is enabled, the dashboard indicates if datasets contain sensitive data.\n- **Findings**: Assess and manage findings generated by AI security and data security policies.\n- **Sensitive data in Vertex AI datasets**: Assess and manage findings of sensitive data in Vertex AI datasets.\n- **Model Armor Findings**: View a graph of prompts and responses that are scanned by\n Model Armor, with detected issues such as prompt injection\n and sensitive data detection.\n\n | **Note:** If Model Armor is not configured for any of the projects in the organization or a template is not created, the Model Armor widgets in the AI Security dashboard won't show any data. For configuration steps, see [Configure\n | Model Armor](/security-command-center/docs/configure-ai-protection#configure-model)"]]