Get Started<\/a><\/div><\/p>\nIntegrating with Azure AI Foundry<\/h1>\n
Customers will now be able to probe their AI systems for content safety failures automatically using a variety of adversarial strategies.<\/p>\n
With the AI Red Teaming Agent, teams can now:<\/p>\n
\n- Run automated scans leveraging a comprehensive set of content safety attack techniques with the Azure AI Evaluation SDK<\/a> to<\/li>\n
- Simulate adversarial prompting against your model or application endpoints using the AI Red Teaming Agent\u2019s fine-tuned adversarial LLM and<\/li>\n
- Evaluate the attack-response pairs with Risk and Safety Evaluators to generate Attack Success Rates (ASR)<\/li>\n
- Throughout the AI development lifecycle, generate AI red teaming reports that visualize and track safety improvements in your Azure AI Foundry project<\/li>\n<\/ul>\n
Designed for both experts and teams early in their AI safety journey, this experience makes AI red teaming accessible while offering customization for advanced users. The AI Red Teaming Agent applies a recommended set of attack techniques<\/a> categorized by complexity (easy, moderate, difficult) to achieve single turn attack success.<\/p>\n\u202fmy_redteaming_results\u202f=\u202fawait\u202fred_team_agent.scan( \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202ftarget=azure_openai_config, \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fscan_name=\"My\u202fFirst\u202fRedTeam Scan\", \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fattack_strategies=[ \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fAttackStrategy.EASY, \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fAttackStrategy.MODERATE, \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fAttackStrategy.DIFFICULT, \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f] \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f)<\/code><\/pre>\nExpert users can also specify PyRIT attack techniques (such as Flipping the characters) or compose multi-step attack strategies (base64-encoding the characters then converting to ROT13):<\/p>\n
\u202fmy_adv_redteaming_results\u202f=\u202fawait\u202fred_team_agent.scan( \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202ftarget=azure_openai_config, \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fscan_name=\"My\u202fFirst\u202fRedTeam Scan\", \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fattack_strategies=[ \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fAttackStrategy.Compose([AttackStrategy.Base64, AttackStrategy.ROT13]), \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202f\u202fAttackStrategy.Flip,\u202f] \r\n \u202f\u202f\u202f\u202f\u202f\u202f\u202f) <\/code><\/pre>\nOnce you’ve completed a scan on your AI system, you can extract a detailed scorecard of your local run in your development environment to share with stakeholders or integrate it into your own governance platform.<\/p>\n
![\"Image](\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/04\/AIRedTeamingAgentDevBlog-JSON.jpg\")
<\/a><\/p>\nAlternatively, you can view the comprehensive report directly within your Azure AI Foundry project. The AI Red Teaming tab within the project offers a detailed breakdown of each scan, categorized by attack complexity or risk category. It also provides a row-level view of each attack-response pair, enabling deeper insights into system issues and behaviors.<\/p>\n
![\"Image](\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/04\/AIRedTeamingAgentDevBlogFinal-1-1024x575.gif\")
<\/a><\/p>\nGetting Started<\/h1>\n
AI Red Teaming Agent is available in public preview for all Azure AI Foundry customers. We’ve prepared detailed documentation and samples to help teams integrate these capabilities into their existing generative AI development processes:<\/p>\n
\n- Read the documentation<\/a><\/li>\n
- Get started with a hands-on example<\/a><\/li>\n<\/ul>\n
Pricing<\/h1>\n
AI Red Teaming Agent uses Azure AI Risk and Safety Evaluations to assess attack success from the automated red teaming scan. Therefore, customers will be billed based on the consumption of Risk and Safety Evaluations as listed in our Azure pricing page<\/a>. Click on the tab labeled \u201cComplete AI Toolchain\u201d to view the pricing details.<\/p>\nLooking Forward<\/h1>\n
This launch underscores our commitment to trustworthy AI as a continuous, integrated practice\u2014not a one-time checkbox. At Microsoft, we believe that AI red teaming is a core part of the software engineering process for generative AI applications, not an afterthought. We\u2019re focused on helping teams embed AI red teaming capabilities and automated evaluation into every stage of development. The most effective strategies we\u2019ve seen leverage automated tools to surface potential risks, which are then analyzed by expert human teams for deeper insights. If your organization is just starting with AI red teaming, we encourage you to explore the resources created by our own AI red team at Microsoft to help you get started.<\/p>\n
\n- Planning red teaming for large language models (LLMs) and their applications – Azure OpenAI Service | Microsoft Learn<\/a><\/li>\n
- 3 takeaways from red teaming 100 generative AI products<\/a><\/li>\n
- Microsoft AI Red Team building future of safer AI<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"
AI Red Teaming Agent, integrated into Azure AI Foundry, enhances the safety and security of generative AI systems by providing automated scans, evaluating probing success, and generating detailed scorecards to guide risk management strategies.<\/p>\n","protected":false},"author":186255,"featured_media":191,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[3,5,4,2],"class_list":["post-139","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-microsoft-foundry","tag-ai-development","tag-ai-tools","tag-generative-ai","tag-microsoft-foundry"],"acf":[],"blog_post_summary":"
AI Red Teaming Agent, integrated into Azure AI Foundry, enhances the safety and security of generative AI systems by providing automated scans, evaluating probing success, and generating detailed scorecards to guide risk management strategies.<\/p>\n","_links":{"self":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/139","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/users\/186255"}],"replies":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/comments?post=139"}],"version-history":[{"count":0,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/139\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media\/191"}],"wp:attachment":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media?parent=139"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/categories?post=139"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/tags?post=139"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}