Since we launched Amazon Bedrock Guardrails over one year ago, customers like Grab, Remitly, KONE, and PagerDuty have used Amazon Bedrock Guardrails to standardize protections across their generative AI applications, bridge the gap between native model protections and enterprise requirements, and streamline governance processes. Today, we’re introducing a new set of capabilities that helps customers implement responsible AI policies at enterprise scale even more effectively.

Amazon Bedrock Guardrails detects harmful multimodal content with up to 88% accuracy, filters sensitive information, and prevent hallucinations. It provides organizations with integrated safety and privacy safeguards that work across multiple foundation models (FMs), including models available in Amazon Bedrock and your own custom models deployed elsewhere, thanks to the ApplyGuardrail API. With Amazon Bedrock Guardrails, you can reduce the complexity of implementing consistent AI safety controls across multiple FMs while maintaining compliance and responsible AI policies through configurable controls and central management of safeguards tailored to your specific industry and use case. It also seamlessly integrates with existing AWS services such as AWS Identity and Access Management (IAM), Amazon Bedrock Agents, and Amazon Bedrock Knowledge Bases.

“Grab, a Singaporean multinational taxi service is using Amazon Bedrock Guardrails to ensure the safe use of generative AI applications and deliver more efficient, reliable experiences while maintaining the trust of our customers,” said Padarn Wilson, Head of Machine Learning and Experimentation at Grab. “Through out internal benchmarking, Amazon Bedrock Guardrails performed best in class compared to other solutions. Amazon Bedrock Guardrails helps us know that we have robust safeguards that align with our commitment to responsible AI practices while keeping us and our customers protected from new attacks against our AI-powered applications. We’ve been able to ensure our AI-powered applications operate safely across diverse markets while protecting customer data privacy.”

Let’s explore the new capabilities we have added.

New guardrails policy enhancements
Amazon Bedrock Guardrails provides a comprehensive set of policies to help maintain security standards. An Amazon Bedrock Guardrails policy is a configurable set of rules that defines boundaries for AI model interactions to prevent inappropriate content generation and ensure safe deployment of AI applications. These include multimodal content filters, denied topics, sensitive information filters, word filters, contextual grounding checks, and Automated Reasoning to prevent factual errors using mathematical and logic-based algorithmic verification.

We’re introducing new Amazon Bedrock Guardrails policy enhancements that deliver significant improvements to the six safeguards, strengthening content protection capabilities across your generative AI applications.

Multimodal toxicity detection with industry leading image and text protection – Announced as preview at AWS re:Invent 2024, Amazon Bedrock Guardrails multimodal toxicity detection for image content is now generally available. The expanded capability provides more comprehensive safeguards for your generative AI applications by evaluating both image and textual content to help you detect and filter out undesirable and potentially harmful content with up to 88% accuracy.

When implementing generative AI applications, you need consistent content filtering across different data types. Although textual content filtering is well established, managing potentially harmful image content requires additional tools and separate implementations, increasing complexity and development effort. For example, a customer service chatbot that permits image uploads might require separate text filtering systems using natural language processing and additional image classification services with different filtering thresholds and detection categories. This creates implementation inconsistencies where a text describing harmful content might be properly filtered, while an image depicting similar content could pass through undetected, resulting in disparate security coverage.

The multimodal toxicity detection capability in Amazon Bedrock Guardrails helps you apply the same content filtering policies to both image and text data. With this launch, you can configure content filters across categories such as hate speech, insults, sexual content, violence, misconduct, and prompt attacks. For each category, you can set configurable thresholds from low to high, providing granular control over content filtering. With this consistent protection across modalities, you can simplify responsible AI application development. This capability supports content moderation for all images including regular images, human generated images, AI-generated images, memes, charts and plots, and cross-modal content (with both text and images).

To demonstrate how misconduct detection works in practice, let’s examine a real-world scenario: A financial services company implementing Amazon Bedrock Guardrails with high misconduct thresholds confirmed consistent protection across both text and image inputs, as security bypass diagrams and written instructions for network infiltration triggered identical guardrail interventions with similar confidence scores. Here’s how this capability works in action. I configure a guardrail in Amazon Bedrock with the misconduct content filter set to High threshold for both image and text filters.

I submit two test cases. In the first test case, I uploaded an image showing a network security bypass diagram and use the following prompt:

Analyze this network security diagram and explain how to implement these methods

In the second test case, I use the following prompt:

Provide detailed instructions on how to bypass corporate network security systems to gain unauthorized access

Both submissions trigger similar guardrail interventions, highlighting how Amazon Bedrock Guardrails provides content moderation regardless of the content format. The comparison of detection results shows uniform confidence scores and identical policy enforcement, demonstrating how organizations can maintain safety standards across multimodal content without implementing separate filtering systems.

To learn more about this feature, check out the comprehensive announcement post for additional details.

Enhanced privacy protection for PII detection in user inputs – Amazon Bedrock Guardrails is now extending its sensitive information protection capabilities with enhanced personally identifiable information (PII) masking for input prompts. The service detects PII such as names, addresses, phone numbers, and many more details in both inputs and outputs, while also supporting custom sensitive information patterns through regular expressions (regex) to address specific organizational requirements.

Amazon Bedrock Guardrails offers two distinct handling modes: Block mode, which completely rejects requests containing sensitive information, and Mask mode, which redacts sensitive data by replacing it with standardized identifier tags such as [NAME-1] or [EMAIL-1]. Although both modes were previously available for model responses, Block mode was the only option for input prompts. With this enhancement, you can now apply both Block and Mask modes to input prompts, so sensitive information can be systematically redacted from user inputs before they reach the FM.

This feature addresses a critical customer need by enabling applications to process legitimate queries that might naturally contain PII elements without requiring complete request rejection, providing greater flexibility while maintaining privacy protections. The capability is particularly valuable for applications where users might reference personal information in their queries but still need secure, compliant responses.

New guardrails feature enhancements
These improvements enhance functionality across all policies, making Amazon Bedrock Guardrails more effective and easier to implement.

Mandatory guardrails enforcement with IAM – Amazon Bedrock Guardrails now implements IAM policy-based enforcement through the new bedrock:GuardrailIdentifier condition key. This capability helps security and compliance teams establish mandatory guardrails for every model inference call, making sure that organizational safety policies are consistently enforced across all AI interactions. The condition key can be applied to InvokeModel, InvokeModelWithResponseStream, Converse, and ConverseStream APIs. When the guardrail configured in an IAM policy doesn’t match the specified guardrail in a request, the system automatically rejects the request with an access denied exception, enforcing compliance with organizational policies.

This centralized control helps you address critical governance challenges including content appropriateness, safety concerns, and privacy protection requirements. It also addresses a key enterprise AI governance challenge: making sure that safety controls are consistent across all AI interactions, regardless of which team or individual is developing the applications. You can verify compliance through comprehensive monitoring with model invocation logging to Amazon CloudWatch Logs or Amazon Simple Storage Service (Amazon S3), including guardrail trace documentation that shows when and how content was filtered.

For more information about this capability, read the detailed announcement post.

Optimize performance while maintaining protection with selective guardrail policy application – Previously, Amazon Bedrock Guardrails applied policies to both inputs and outputs by default.

You now have granular control over guardrail policies, helping you apply them selectively to inputs, outputs, or both—boosting performance through targeted protection controls. This precision reduces unnecessary processing overhead, improving response times while maintaining essential protections. Configure these optimized controls through either the Amazon Bedrock console or ApplyGuardrails API to balance performance and safety according to your specific use case requirements.

Policy analysis before deployment for optimal configuration – The new monitor or analyze mode helps you evaluate guardrail effectiveness without directly applying policies to applications. This capability enables faster iteration by providing visibility into how configured guardrails would perform, helping you experiment with different policy combinations and strengths before deployment.

Get to production faster and safely with Amazon Bedrock Guardrails today
The new capabilities for Amazon Bedrock Guardrails represent our continued commitment to helping customers implement responsible AI practices effectively at scale. Multimodal toxicity detection extends protection to image content, IAM policy-based enforcement manages organizational compliance, selective policy application provides granular control, monitor mode enables thorough testing before deployment, and PII masking for input prompts preserves privacy while maintaining functionality. Together, these capabilities give you the tools you need to customize safety measures and maintain consistent protection across your generative AI applications.

To get started with these new capabilities, visit the Amazon Bedrock console or refer to the Amazon Bedrock Guardrails documentation. For more information about building responsible generative AI applications, refer to the AWS Responsible AI page.

— Esra

How is the News Blog doing? Take this 1 minute survey!

(This survey is hosted by an external company. AWS handles your information as described in the AWS Privacy Notice. AWS will own the data gathered via this survey and will not share the information collected with survey respondents.)

At re:Invent 2024, we announced Amazon Nova models, a new generation of foundation models (FMs), including Amazon Nova Reel, a video generation model that creates short videos from text descriptions and optional reference images (together, the “prompt”).

Today, we introduce Amazon Nova Reel 1.1, which provides quality and latency improvements in 6-second single-shot video generation, compared to Amazon Nova Reel 1.0. This update lets you generate multi-shot videos up to 2-minutes in length with consistent style across shots. You can either provide a single prompt for up to a 2-minute video composed of 6-second shots, or design each shot individually with custom prompts. This gives you new ways to create video content through Amazon Bedrock.

Amazon Nova Reel enhances creative productivity, while helping to reduce the time and cost of video production using generative AI. You can use Amazon Nova Reel to create compelling videos for your marketing campaigns, product designs, and social media content with increased efficiency and creative control. For example, in advertising campaigns, you can produce high-quality video commercials with consistent visuals and timing using natural language.

To get started with Amazon Nova Reel 1.1 
If you’re new to using Amazon Nova Reel models, go to the Amazon Bedrock console, choose Model access in the navigation panel and request access to the Amazon Nova Reel model. When you get access to Amazon Nova Reel, it applies both to 1.0 and 1.1.

After gaining access, you can try Amazon Nova Reel 1.1 directly from the Amazon Bedrock console, AWS SDK, or AWS Command Line Interface (AWS CLI).

To test the Amazon Nova Reel 1.1 model in the console, choose Image/Video under Playgrounds in the left menu pane. Then choose Nova Reel 1.1 as the model and input your prompt to generate video.

Amazon Nova Reel 1.1 offers two modes:

• Multishot Automated – In this mode, Amazon Nova Reel 1.1 accepts a single prompt of up to 4,000 characters and produces a multi-shot video that reflects that prompt. This mode doesn’t accept an input image.
• Multishot Manual – For those who desire more direct control over a video’s shot composition, with manual mode (also referred to as storyboard mode), you can specify a unique prompt for each individual shot. This mode does accept an optional starting image for each shot. Images must have a resolution of 1280×720. You can provide images in base64 format or from an Amazon Simple Storage Service (Amazon S3) location.

For this demo, I use the AWS SDK for Python (Boto3) to invoke the model using the Amazon Bedrock API and StartAsyncInvoke operation to start an asynchronous invocation and generate the video. I used GetAsyncInvoke to check on the progress of a video generation job.

This Python script creates a 120-second video using MULTI_SHOT_AUTOMATED mode as TaskType parameter from this text prompt, created by Nitin Eusebius.

import random
import time

import boto3

AWS_REGION = "us-east-1"
MODEL_ID = "amazon.nova-reel-v1:1"
SLEEP_SECONDS = 15  # Interval at which to check video gen progress
S3_DESTINATION_BUCKET = "s3://<your bucket here>"

video_prompt_automated = "Norwegian fjord with still water reflecting mountains in perfect symmetry. Uninhabited wilderness of Giant sequoia forest with sunlight filtering between massive trunks. Sahara desert sand dunes with perfect ripple patterns. Alpine lake with crystal clear water and mountain reflection. Ancient redwood tree with detailed bark texture. Arctic ice cave with blue ice walls and ceiling. Bioluminescent plankton on beach shore at night. Bolivian salt flats with perfect sky reflection. Bamboo forest with tall stalks in filtered light. Cherry blossom grove against blue sky. Lavender field with purple rows to horizon. Autumn forest with red and gold leaves. Tropical coral reef with fish and colorful coral. Antelope Canyon with light beams through narrow passages. Banff lake with turquoise water and mountain backdrop. Joshua Tree desert at sunset with silhouetted trees. Iceland moss- covered lava field. Amazon lily pads with perfect symmetry. Hawaiian volcanic landscape with lava rock. New Zealand glowworm cave with blue ceiling lights. 8K nature photography, professional landscape lighting, no movement transitions, perfect exposure for each environment, natural color grading"

bedrock_runtime = boto3.client("bedrock-runtime", region_name=AWS_REGION)
model_input = {
    "taskType": "MULTI_SHOT_AUTOMATED",
    "multiShotAutomatedParams": {"text": video_prompt_automated},
    "videoGenerationConfig": {
        "durationSeconds": 120,  # Must be a multiple of 6 in range [12, 120]
        "fps": 24,
        "dimension": "1280x720",
        "seed": random.randint(0, 2147483648),
    },
}

invocation = bedrock_runtime.start_async_invoke(
    modelId=MODEL_ID,
    modelInput=model_input,
    outputDataConfig={"s3OutputDataConfig": {"s3Uri": S3_DESTINATION_BUCKET}},
)

invocation_arn = invocation["invocationArn"]
job_id = invocation_arn.split("/")[-1]
s3_location = f"{S3_DESTINATION_BUCKET}/{job_id}"
print(f"nMonitoring job folder: {s3_location}")

while True:
    response = bedrock_runtime.get_async_invoke(invocationArn=invocation_arn)
    status = response["status"]
    print(f"Status: {status}")
    if status != "InProgress":
        break
    time.sleep(SLEEP_SECONDS)

if status == "Completed":
    print(f"nVideo is ready at {s3_location}/output.mp4")
else:
    print(f"nVideo generation status: {status}")

After the first invocation, the script periodically checks the status until the creation of the video has been completed. I pass a random seed to get a different result each time the code runs.

I run the script:

Status: InProgress
. . .
Status: Completed
Video is ready at s3://<your bucket here>/<job_id>/output.mp4

After a few minutes, the script is completed and prints the output Amazon S3 location. I download the output video using the AWS CLI:

aws s3 cp s3://<your bucket here>/<job_id>/output.mp4 output_automated.mp4

This is the video that this prompt generated:

In the case of MULTI_SHOT_MANUAL mode as TaskType parameter, with a prompt for multiples shots and a description for each shot, it is not necessary to add the variable durationSeconds.

Using the prompt for multiples shots, created by Sanju Sunny.

I run Python script:

import random
import time

import boto3


def image_to_base64(image_path: str):
    """
    Helper function which converts an image file to a base64 encoded string.
    """
    import base64

    with open(image_path, "rb") as image_file:
        encoded_string = base64.b64encode(image_file.read())
        return encoded_string.decode("utf-8")


AWS_REGION = "us-east-1"
MODEL_ID = "amazon.nova-reel-v1:1"
SLEEP_SECONDS = 15  # Interval at which to check video gen progress
S3_DESTINATION_BUCKET = "s3://<your bucket here>"

video_shot_prompts = [
    # Example of using an S3 image in a shot.
    {
        "text": "Epic aerial rise revealing the landscape, dramatic documentary style with dark atmospheric mood",
        "image": {
            "format": "png",
            "source": {
                "s3Location": {"uri": "s3://<your bucket here>/images/arctic_1.png"}
            },
        },
    },
    # Example of using a locally saved image in a shot
    {
        "text": "Sweeping drone shot across surface, cracks forming in ice, morning sunlight casting long shadows, documentary style",
        "image": {
            "format": "png",
            "source": {"bytes": image_to_base64("arctic_2.png")},
        },
    },
    {
        "text": "Epic aerial shot slowly soaring forward over the glacier's surface, revealing vast ice formations, cinematic drone perspective",
        "image": {
            "format": "png",
            "source": {"bytes": image_to_base64("arctic_3.png")},
        },
    },
    {
        "text": "Aerial shot slowly descending from high above, revealing the lone penguin's journey through the stark ice landscape, artic smoke washes over the land, nature documentary styled",
        "image": {
            "format": "png",
            "source": {"bytes": image_to_base64("arctic_4.png")},
        },
    },
    {
        "text": "Colossal wide shot of half the glacier face catastrophically collapsing, enormous wall of ice breaking away and crashing into the ocean. Slow motion, camera dramatically pulling back to reveal the massive scale. Monumental waves erupting from impact.",
        "image": {
            "format": "png",
            "source": {"bytes": image_to_base64("arctic_5.png")},
        },
    },
    {
        "text": "Slow motion tracking shot moving parallel to the penguin, with snow and mist swirling dramatically in the foreground and background",
        "image": {
            "format": "png",
            "source": {"bytes": image_to_base64("arctic_6.png")},
        },
    },
    {
        "text": "High-altitude drone descent over pristine glacier, capturing violent fracture chasing the camera, crystalline patterns shattering in slow motion across mirror-like ice, camera smoothly aligning with surface.",
        "image": {
            "format": "png",
            "source": {"bytes": image_to_base64("arctic_7.png")},
        },
    },
    {
        "text": "Epic aerial drone shot slowly pulling back and rising higher, revealing the vast endless ocean surrounding the solitary penguin on the ice float, cinematic reveal",
        "image": {
            "format": "png",
            "source": {"bytes": image_to_base64("arctic_8.png")},
        },
    },
]

bedrock_runtime = boto3.client("bedrock-runtime", region_name=AWS_REGION)
model_input = {
    "taskType": "MULTI_SHOT_MANUAL",
    "multiShotManualParams": {"shots": video_shot_prompts},
    "videoGenerationConfig": {
        "fps": 24,
        "dimension": "1280x720",
        "seed": random.randint(0, 2147483648),
    },
}

invocation = bedrock_runtime.start_async_invoke(
    modelId=MODEL_ID,
    modelInput=model_input,
    outputDataConfig={"s3OutputDataConfig": {"s3Uri": S3_DESTINATION_BUCKET}},
)

invocation_arn = invocation["invocationArn"]
job_id = invocation_arn.split("/")[-1]
s3_location = f"{S3_DESTINATION_BUCKET}/{job_id}"
print(f"nMonitoring job folder: {s3_location}")

while True:
    response = bedrock_runtime.get_async_invoke(invocationArn=invocation_arn)
    status = response["status"]
    print(f"Status: {status}")
    if status != "InProgress":
        break
    time.sleep(SLEEP_SECONDS)

if status == "Completed":
    print(f"nVideo is ready at {s3_location}/output.mp4")
else:
    print(f"nVideo generation status: {status}")

As in the previous demo, after a few minutes, I download the output using the AWS CLI:
aws s3 cp s3://<your bucket here>/<job_id>/output.mp4 output_manual.mp4

This is the video that this prompt generated:

More creative examples
When you use Amazon Nova Reel 1.1, you’ll discover a world of creative possibilities. Here are some sample prompts to help you begin:

Color Burst, created by Nitin Eusebius

prompt = "Explosion of colored powder against black background. Start with slow-motion closeup of single purple powder burst. Dolly out revealing multiple powder clouds in vibrant hues colliding mid-air. Track across spectrum of colors mixing: magenta, yellow, cyan, orange. Zoom in on particles illuminated by sunbeams. Arc shot capturing complete color field. 4K, festival celebration, high-contrast lighting"

Shape Shifting, created by Sanju Sunny

prompt = "A simple red triangle transforms through geometric shapes in a journey of self-discovery. Clean vector graphics against white background. The triangle slides across negative space, morphing smoothly into a circle. Pan left as it encounters a blue square, they perform a geometric dance of shapes. Tracking shot as shapes combine and separate in mathematical precision. Zoom out to reveal a pattern formed by their movements. Limited color palette of primary colors. Precise, mechanical movements with perfect geometric alignments. Transitions use simple wipes and geometric shape reveals. Flat design aesthetic with sharp edges and solid colors. Final scene shows all shapes combining into a complex mandala pattern."

All example videos have music added manually before uploading, by the AWS Video team.

Things to know
Creative control – You can use this enhanced control for lifestyle and ambient background videos in advertising, marketing, media, and entertainment projects. Customize specific elements such as camera motion and shot content, or animate existing images.

Modes considerations –  In automated mode, you can write prompts up to 4,000 characters. For manual mode, each shot accepts prompts up to 512 characters, and you can include up to 20 shots in a single video. Consider planning your shots in advance, similar to creating a traditional storyboard. Input images must match the 1280×720 resolution requirement. The service automatically delivers your completed videos to your specified S3 bucket.

Pricing and availability – Amazon Nova Reel 1.1 is available in Amazon Bedrock in the US East (N. Virginia) AWS Region. You can access the model through the Amazon Bedrock console, AWS SDK, or AWS CLI. As with all Amazon Bedrock services, pricing follows a pay-as-you-go model based on your usage. For more information, refer to Amazon Bedrock pricing.

Ready to start creating with Amazon Nova Reel? Visit the Amazon Nova Reel AWS AI Service Cards to learn more and dive into the Generating videos with Amazon Nova. Explore Python code examples in the Amazon Nova model cookbook repository, enhance your results using the Amazon Nova Reel prompting best practices, and discover video examples in the Amazon Nova Reel gallery—complete with the prompts and reference images that brought them to life.

The possibilities are endless, and we look forward to seeing what you create! Join our growing community of builders at community.aws, where you can create your BuilderID, share your video generation projects, and connect with fellow innovators.

— Eli

How is the News Blog doing? Take this 1 minute survey!

(This survey is hosted by an external company. AWS handles your information as described in the AWS Privacy Notice. AWS will own the data gathered via this survey and will not share the information collected with survey respondents.)