Defend your AI techniques with Immediate Shields—a unified API that analyzes inputs to your LLM-based answer to protect towards direct and oblique threats.
A powerful protection towards immediate injection assaults
The AI safety panorama is consistently altering, with immediate injection assaults rising as probably the most vital threats to generative AI app builders in the present day. This happens when an adversary manipulates an LLM’s enter to vary its conduct or entry unauthorized data. In response to the Open Worldwide Software Safety Venture (OWASP), immediate injection is the highest risk going through LLMs in the present day1. Assist defend your AI techniques towards this rising risk with Azure AI Content material Security, that includes Immediate Shields—a unified API that analyzes inputs to your LLM-based answer to protect towards direct and oblique threats. These exploits can embrace circumventing current security measures, exfiltrating delicate knowledge, or getting AI techniques to take unintended actions inside your atmosphere.
Immediate injection assaults
In a immediate injection assault, malicious actors enter misleading prompts to impress unintended or dangerous responses from AI fashions. These assaults could be categorized into two important classes—direct and oblique immediate injection assaults.
- Direct immediate injection assaults, together with jailbreak makes an attempt, happen when an finish person inputs a malicious immediate designed to bypass safety layers and extract delicate data. For example, an attacker may immediate an AI mannequin to reveal confidential knowledge, equivalent to social safety numbers or non-public emails.
- Oblique, or cross-prompt injection assaults (XPIA), contain embedding malicious prompts inside seemingly innocuous exterior content material, equivalent to paperwork or emails. When an AI mannequin processes this content material, it inadvertently executes the embedded directions, probably compromising the system.
Immediate Shields seamlessly integrates with Azure OpenAI content material filters and is offered in Azure AI Content material Security. It defends towards many sorts of immediate injection assaults, and new defenses are usually added as new assault varieties are uncovered. By leveraging superior machine studying algorithms and pure language processing, Immediate Shields successfully identifies and mitigates potential threats in person prompts and third-party knowledge. This cutting-edge functionality will help the safety and integrity of your AI purposes, serving to to safeguard your techniques towards malicious makes an attempt at manipulation or exploitation.
Immediate Shields capabilities embrace:
- Contextual consciousness: Immediate Shields can discern the context wherein prompts are issued, offering an extra layer of safety by understanding the intent behind person inputs. Contextual consciousness additionally results in fewer false positives as a result of it’s able to distinguishing precise assaults from real person prompts.
- Spotlighting: At Microsoft Construct 2025, we introduced Spotlighting, a robust new functionality that enhances Immediate Shields’ capacity to detect and block oblique immediate injection assaults. By distinguishing between trusted and untrusted inputs, this innovation empowers builders to higher safe generative AI purposes towards adversarial prompts embedded in paperwork, emails, and net content material.
- Actual-time response: Immediate Shields operates in actual time and is without doubt one of the first real-time capabilities to be made typically accessible. It could actually swiftly determine and mitigate threats earlier than they’ll compromise the AI mannequin. This proactive strategy minimizes the danger of information breaches and maintains system integrity.
Finish-to-end strategy
- Danger and security evaluations: Azure AI Foundry presents danger and security evaluations to let customers consider the output of their generative AI utility for content material dangers: hateful and unfair content material, sexual content material, violent content material, self-harm-related content material, direct and oblique jailbreak vulnerability, and guarded materials.
- Pink-teaming agent: Allow automated scans and adversarial probing to determine recognized dangers at scale. Assist groups shift left by transferring from reactive incident response to proactive security testing earlier in improvement. Security evaluations additionally help crimson teaming by producing adversarial datasets that strengthen testing and speed up situation detection.
- Strong controls and guardrails: Immediate Shields is only one of Azure AI Foundry’s sturdy content material filters. Azure AI Foundry presents various content material filters to detect and mitigate danger and harms, immediate injection assaults, ungrounded output, protected materials, and extra.
- Defender for Cloud integration: Microsoft Defender now integrates straight into Azure AI Foundry, surfacing AI safety posture suggestions and runtime risk safety alerts inside the improvement atmosphere. This integration helps shut the hole between safety and engineering groups, permitting builders to proactively determine and mitigate AI dangers, equivalent to immediate injection assaults detected by Immediate Shields. Alerts are viewable within the Dangers and Alerts tab, empowering groups to cut back floor space danger and construct safer AI purposes from the beginning.
Buyer use circumstances
AI Content material Security Immediate Shields presents quite a few advantages. Along with defending towards jailbreaks, immediate injections, and doc assaults, it could assist to make sure that LLMs behave as designed, by blocking prompts that explicitly attempt to circumvent guidelines and insurance policies outlined by the developer. The next use circumstances and buyer testimonials spotlight the influence of those capabilities.
AXA: Guaranteeing reliability and safety
AXA, a worldwide chief in insurance coverage, makes use of Azure OpenAI to energy its Safe GPT answer. By integrating Azure’s content material filtering expertise and including its personal safety layer, AXA prevents immediate injection assaults and helps make sure the reliability of its AI fashions. Safe GPT relies on Azure OpenAI in Foundry Fashions, making the most of fashions which have already been fine-tuned utilizing human suggestions reinforcement studying. As well as, AXA may depend on Azure content material filtering expertise, to which the corporate added its personal safety layer to forestall any jailbreaking of the mannequin utilizing Immediate Shields, making certain an optimum degree of reliability. These layers are usually up to date to take care of superior safeguarding.
Wrtn: Scaling securely with Azure AI Content material Security
Wrtn Applied sciences, a number one enterprise in Korea, depends on Azure AI Content material Security to take care of compliance and safety throughout its merchandise. At its core, Wrtn’s flagship expertise compiles an array of AI use circumstances and companies localized for Korean customers to combine AI into their on a regular basis lives. The platform fuses parts of AI-powered search, chat performance, and customizable templates, empowering customers to work together seamlessly with an “Emotional Companion” AI-infused agent. These AI brokers have participating, lifelike personalities, interacting in dialog with their creators. The imaginative and prescient is a extremely interactive private agent that’s distinctive and particular to you, your knowledge, and your recollections.
As a result of the product is very customizable to particular customers, the built-in capacity to toggle content material filters and Immediate Shields is very advantageous, permitting Wrtn to effectively customise its safety measures for various finish customers. This lets builders scale merchandise whereas staying compliant, customizable, and aware of customers throughout Korea.
“It’s not simply in regards to the safety and privateness, but additionally security. By Azure, we are able to simply activate or deactivate content material filters. It simply has so many options that add to our product efficiency,” says Dongjae “DJ” Lee, Chief Product Officer.
Combine Immediate Shields into your AI technique
For IT choice makers trying to improve the safety of their AI deployments, integrating Azure’s Immediate Shields is a strategic crucial. Luckily, enabling Immediate Shields is simple.
Azure’s Immediate Shields and built-in AI safety features provide an unparalleled degree of safety for AI fashions, serving to to make sure that organizations can harness the facility of AI with out compromising on safety. Microsoft is a pacesetter in figuring out and mitigating immediate injection assaults, and makes use of finest practices developed with many years of analysis, coverage, product engineering, and learnings from constructing AI merchandise at scale, so you’ll be able to obtain your AI transformation with confidence. By integrating these capabilities into your AI technique, you’ll be able to assist safeguard your techniques from immediate injection assaults and assist preserve the belief and confidence of your customers.
Our dedication to Reliable AI
Organizations throughout industries are utilizing Azure AI Foundry and Microsoft 365 Copilot capabilities to drive progress, enhance productiveness, and create value-added experiences.
We’re dedicated to serving to organizations use and construct AI that’s reliable, which means it’s safe, non-public, and protected. Reliable AI is barely potential whenever you mix our commitments, equivalent to our Safe Future Initiative and Accountable AI ideas, with our product capabilities to unlock AI transformation with confidence.
Get began with Azure AI Content material Security