[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-02-05。"],[[["\u003cp\u003eAI significantly enhances cloud security by reducing threats, minimizing manual effort, and addressing the shortage of cybersecurity experts.\u003c/p\u003e\n"],["\u003cp\u003eAI capabilities, including Gemini in Security and built-in Google Cloud services, can assist at every security lifecycle stage, from analyzing malicious code to summarizing alerts.\u003c/p\u003e\n"],["\u003cp\u003eUsing AI for security can achieve increasing levels of autonomy: manual, assisted, semi-autonomous, and autonomous, where AI increasingly leads security tasks.\u003c/p\u003e\n"],["\u003cp\u003eAI helps simplify security, making it more accessible to experts and non-experts alike, by summarizing alerts, recommending mitigations, and automating tasks like malware analysis.\u003c/p\u003e\n"],["\u003cp\u003eImplementing secure development practices for AI systems is crucial, and AI can be leveraged for secure coding, data cleaning, and validating tools, which are relevant to all levels of security autonomy.\u003c/p\u003e\n"]]],[],null,["# Use AI for security\n\nThis principle in the security pillar of the\n[Google Cloud Well-Architected Framework](/architecture/framework)\nprovides recommendations to use AI to help you improve the security of your\ncloud workloads.\n\nBecause of the increasing number and sophistication of cyber attacks, it's\nimportant to take advantage of AI's potential to help improve security. AI can\nhelp to reduce the number of threats, reduce the manual effort required by security\nprofessionals, and help compensate for the scarcity of experts in the cyber-security\ndomain.\n\nPrinciple overview\n------------------\n\nUse AI capabilities to improve your existing security systems and processes.\nYou can use\n[Gemini in Security](/security/ai)\nas well as the intrinsic AI capabilities that are built into Google Cloud services.\n\nThese AI capabilities can transform security by providing assistance across\nevery stage of the security lifecycle. For example, you can use AI to do the\nfollowing:\n\n- Analyze and explain potentially malicious code without reverse engineering.\n- Reduce repetitive work for cyber-security practitioners.\n- Use natural language to generate queries and interact with security event data.\n- Surface contextual information.\n- Offer recommendations for quick responses.\n- Aid in the remediation of events.\n- Summarize high-priority alerts for misconfigurations and vulnerabilities, highlight potential impacts, and recommend mitigations.\n\nLevels of security autonomy\n---------------------------\n\nAI and automation can help you achieve better security outcomes when you're\ndealing with ever-evolving cyber-security threats. By using AI for security, you\ncan achieve greater levels of autonomy to detect and prevent threats and improve\nyour overall security posture. Google defines four\n[levels of autonomy](https://services.google.com/fh/files/misc/google-cloud-product-vision-ai-powered-security.pdf#page=6)\nwhen you use AI for security, and they outline the increasing role of AI in\nassisting and eventually leading security tasks:\n\n1. **Manual**: Humans run all of the security tasks (prevent, detect, prioritize, and respond) across the entire security lifecycle.\n2. **Assisted**: AI tools, like Gemini, boost human productivity by summarizing information, generating insights, and making recommendations.\n3. **Semi-autonomous**: AI takes primary responsibility for many security tasks and delegates to humans only when required.\n4. **Autonomous**: AI acts as a trusted assistant that drives the security lifecycle based on your organization's goals and preferences, with minimal human intervention.\n\nRecommendations\n---------------\n\nThe following sections describe the recommendations for using AI for security.\nThe sections also indicate how the recommendations align with Google's Secure AI\nFramework (SAIF)\n[core elements](https://developers.google.com/machine-learning/resources/saif)\nand how they're relevant to the\n[levels of security autonomy](#levels_of_security_autonomy).\n\n- [Enhance threat detection and response with AI](#enhance_threat_detection_and_response_with_ai)\n- [Simplify security for experts and non-experts](#simplify_security_for_experts_and_non-experts)\n- [Automate time-consuming security tasks with AI](#automate_time-consuming_security_tasks_with_ai)\n- [Incorporate AI into risk management and governance processes](#incorporate_ai_into_risk_management_and_governance_processes)\n- [Implement secure development practices for AI systems](#implement_secure_development_practices_for_ai_systems)\n\n| **Note:** For more information about Google Cloud's overall vision for using Gemini across our products to accelerate AI for security, see the whitepaper [Google Cloud's Product Vision for AI-Powered Security](https://services.google.com/fh/files/misc/google-cloud-product-vision-ai-powered-security.pdf).\n\n### Enhance threat detection and response with AI\n\nThis recommendation is relevant to the following\n[focus areas](/architecture/framework/security#focus_areas_of_cloud_security):\n\n- Security operations (SecOps)\n- Logging, auditing, and monitoring\n\nAI can analyze large volumes of security data, offer insights into threat actor\nbehavior, and automate the analysis of potentially malicious code. This\nrecommendation is aligned with the following SAIF elements:\n\n- Extend detection and response to bring AI into your organization's threat universe.\n- Automate defenses to keep pace with existing and new threats.\n\nDepending on your implementation, this recommendation can be relevant to the\nfollowing levels of autonomy:\n\n- **Assisted**: AI helps with threat analysis and detection.\n- **Semi-autonomous**: AI takes on more responsibility for the security task.\n\n[Google Threat Intelligence](/security/products/threat-intelligence),\nwhich uses AI to analyze threat actor behavior and malicious code, can help you\nimplement this recommendation.\n\n### Simplify security for experts and non-experts\n\nThis recommendation is relevant to the following\n[focus areas](/architecture/framework/security#focus_areas_of_cloud_security):\n\n- Security operations (SecOps)\n- Cloud governance, risk, and compliance\n\nAI-powered tools can summarize alerts and recommend mitigations, and these\ncapabilities can make security more accessible to a wider range of personnel.\nThis recommendation is aligned with the following SAIF elements:\n\n- Automate defenses to keep pace with existing and new threats.\n- Harmonize platform-level controls to ensure consistent security across the organization.\n\nDepending on your implementation, this recommendation can be relevant to the\nfollowing levels of autonomy:\n\n- **Assisted**: AI helps you to improve the accessibility of security information.\n- **Semi-autonomous**: AI helps to make security practices more effective for all users.\n\nGemini in\n[Security Command Center](/security/products/security-command-center)\ncan provide summaries of alerts for misconfigurations and vulnerabilities.\n\n### Automate time-consuming security tasks with AI\n\nThis recommendation is relevant to the following\n[focus areas](/architecture/framework/security#focus_areas_of_cloud_security):\n\n- Infrastructure security\n- Security operations (SecOps)\n- Application security\n\nAI can automate tasks such as analyzing malware, generating security rules, and\nidentifying misconfigurations. These capabilities can help to reduce the\nworkload on security teams and accelerate response times. This recommendation is\naligned with the SAIF element about automating defenses to keep pace with\nexisting and new threats.\n\nDepending on your implementation, this recommendation can be relevant to the\nfollowing levels of autonomy:\n\n- **Assisted**: AI helps you to automate tasks.\n- **Semi-autonomous**: AI takes primary responsibility for security tasks, and only requests human assistance when needed.\n\nGemini in\n[Google SecOps](/security/products/security-operations)\ncan help to automate high-toil tasks by assisting analysts, retrieving relevant\ncontext, and making recommendations for next steps.\n\n### Incorporate AI into risk management and governance processes\n\nThis recommendation is relevant to the following\n[focus area](/architecture/framework/security#focus_areas_of_cloud_security):\nCloud governance, risk, and compliance.\n\nYou can use AI to build a model inventory and risk profiles. You can also use AI\nto implement policies for data privacy, cyber risk, and third-party risk. This\nrecommendation is aligned with the SAIF element about contextualizing AI system\nrisks in surrounding business processes.\n\nDepending on your implementation, this recommendation can be relevant to the\nsemi-autonomous level of autonomy. At this level, AI can orchestrate security\nagents that run processes to achieve your custom security goals.\n\n### Implement secure development practices for AI systems\n\nThis recommendation is relevant to the following\n[focus areas](/architecture/framework/security#focus_areas_of_cloud_security):\n\n- Application security\n- AI and ML security\n\nYou can use AI for secure coding, cleaning training data, and validating tools\nand artifacts. This recommendation is aligned with the SAIF element about\nexpanding strong security foundations to the AI ecosystem.\n\nThis recommendation can be relevant to all levels of security autonomy, because\na secure AI system needs to be in place before AI can be used effectively for\nsecurity. The recommendation is most relevant to the assisted level, where\nsecurity practices are augmented by AI.\n\nTo implement this recommendation, follow the\n[Supply-chain Levels for Software Artifacts (SLSA)](https://slsa.dev)\nguidelines for AI artifacts and use validated container images."]]