AWS ai documentation change
Summary
Updated overview documentation for Amazon Nova 2 Sonic with clarifications on model interaction, expanded workflow examples, updated capabilities list (reduced voice count from 22 to 18), restructured sections with markdown formatting changes, added glossary definitions, and detailed updates to privacy and security sections including session data handling, voice cloning protections, and updated links.
Security assessment
The changes include security-related documentation updates such as: 1) Enhanced privacy details about temporary session context storage in memory and garbage collection, 2) Explicit mention of guardrails to prevent voice cloning and impersonation, 3) Clarification that customer data is encrypted and customers can use their own keys via AWS KMS. However, there is no concrete evidence these changes address a specific security vulnerability or incident; they appear to be routine documentation improvements and clarifications of existing security features.
Diff
diff --git a/ai/responsible-ai/nova-2-sonic/overview.md b/ai/responsible-ai/nova-2-sonic/overview.md index 229a2578b..2ab516ad4 100644 --- a//ai/responsible-ai/nova-2-sonic/overview.md +++ b//ai/responsible-ai/nova-2-sonic/overview.md @@ -5 +5 @@ -Overview of Amazon Nova 2 SonicIntended use cases and limitationsDesign of Amazon Nova 2 SonicDeployment and performance optimization best practicesFurther informationGlossary +Overview of Amazon Nova 2 SonicIntended use cases and limitations @@ -21 +21 @@ Amazon Nova 2 Sonic is a proprietary multi-modal foundation model (FM) designed -An Amazon Nova 2 Sonic <text prompt, speech or text inputs, and generated speech and text> triple is said to be "effective" if a skilled human evaluator decides that the resulting conversation is of reasonable quality in terms of 1/ accuracy of speech recognition, 2/ robustness to different acoustic conditions, 3/ expressivity in the generated speech response, 4/ efficiency in dialog handling, and 5/ the relevance, coherence, and consistency of the content of the responses. Otherwise, the output may require refinement or may not fully meet all evaluation criteria for the specific use case. A customer's implementation must decide if a resulting conversation is effective using human judgment, whether human judgement is applied on a case-by-case basis (as happens when the Amazon Bedrock Console playground is used by itself) or is applied via the customer's choice of an acceptable score on an automated test. +Customers invoke the Amazon Nova 2 Sonic model by providing instructions (e.g., text prompt, chat history, tools, etc.) along with a speech or text input. The model uses the instructions and input to generate a speech and text response. Together, this <instructions, input, speech or text input, and generated speech and text response> make up a single interaction turn. We say that a turn is ”effective“ if a skilled human evaluator decides that the resulting conversation is of reasonable quality in terms of 1/ accuracy of speech recognition, 2/ robustness to different acoustic conditions, 3/ expressivity in the generated speech response, 4/ efficiency in dialog handling, and 5/ the relevance, coherence, and consistency of the content of the responses. Otherwise, the output may require refinement or may not fully meet all evaluation criteria for the specific use case. A customer's implementation must decide if a resulting conversation is effective using human judgment, whether human judgement is applied on a case-by-case basis (as happens when the Amazon Bedrock Console playground is used by itself) or is applied via the customer's choice of an acceptable score on an automated test. @@ -23 +23 @@ An Amazon Nova 2 Sonic <text prompt, speech or text inputs, and generated speech -The "overall effectiveness" of any foundation model for a specific use case is based on the percentage of use-case specific inputs for which the model returns an effective result. Customers should define and measure effectiveness for themselves for the following reasons. First, the customer is best positioned to know which triples will best represent their use case, and should therefore be included in an evaluation dataset. Second, different speech-to-speech models may respond differently to the same prompt, requiring tuning of the prompt and/or the evaluation mechanism. +The "overall effectiveness" of any foundation model for a specific use case is based on the percentage of use-case specific inputs for which the model returns an effective result. Customers should define and measure effectiveness for themselves for the following reasons. First, the customer is best positioned to know which turns will best represent their use case and should therefore be included in an evaluation dataset. Second, different speech-to-speech models may respond differently to the same prompt, requiring tuning of the prompt and/or the evaluation mechanism. @@ -29 +29 @@ Like all AI systems, Amazon Nova 2 Sonic is designed to understand relevant diff -Amazon Nova 2 Sonic serves a wide range of potential application domains and offers the following core capabilities: low latency, speech understanding in multiple languages, speech generation in twenty-two expressive voices, natural and efficient dialog handling, cross-modal input, tool use, and asynchronous task completion. +Amazon Nova 2 Sonic is a foundation model that customers can integrate into their own applications and agentic systems. While Amazon Nova 2 Sonic provides core capabilities out-of-the-box, customers are responsible for building the surrounding application infrastructure, including agent frameworks, tool integrations, database connections, and business logic. Amazon Nova 2 Sonic serves a wide range of potential application domains and offers the following core capabilities: low latency, speech understanding in multiple languages, speech generation in eighteen expressive voices, natural and efficient dialog handling, cross-modal input, tool use, and asynchronous task completion. @@ -39 +39 @@ Amazon Nova 2 Sonic serves a wide range of potential application domains and off - * **Cross-modal input** for also taking text messages to converse with the AI agent in the same session, without loss of context. + * **Cross-modal input** that allows the model to process both speech and text inputs, enabling customers to build applications where users can seamlessly switch between speaking and typing. @@ -41 +41 @@ Amazon Nova 2 Sonic serves a wide range of potential application domains and off - * **Tool use** , enabling precise responses based on specific enterprise data and agentic workflows to resolve customer queries or complete specific tasks (for example, making a reservation). + * **Tool use** , allowing customer-built agents to invoke external tools and APIs that access enterprise databases, complete transactions, or perform specific tasks in agentic workflows. @@ -52 +52 @@ When assessing a speech-to-speech model for a particular use case, we encourage -Consider the following use case of utilizing Amazon Nova 2 Sonic to power an AI agent that helps a hotel guest cancel their upcoming reservation through natural voice conversation via a phone call. The **business goal** is to provide an efficient, 24/7 self-service voice assistant that can handle hotel reservation cancellations while maintaining customer satisfaction and reducing call center volume. The **stakeholders** include the hotel guest, who wants to quickly cancel their reservation and receive confirmation without friction, and the hotel customer service team, who wants to automate routine cancellations to focus on more complex guest issues. The **workflow** is: 1/ the guest initiates a voice conversation with the Amazon Nova 2 Sonic-powered hotel assistant; 2/ the system prompts the guest for identification and reservation details; 3/ the guest provides the necessary information through speech; 4/ the system verifies the reservation in the hotel's database; 5/ the system informs the guest of any applicable cancellation policies or fees; 6/ the guest confirms their desire to proceed with the cancellation; 7/ the system processes the cancellation in the hotel's reservation system; and 8/ the system provides verbal confirmation and sends a follow-up email receipt. **Input prompts** contain information regarding reservation identification (confirmation number, booking dates), guest identification (name, contact information), reason for cancellation (optional), confirmation decisions (yes/no responses), and questions about cancellation policies or fees. **Output content** contains information regarding verification of reservation details, information about cancellation policies and any applicable fees, confirmation of successful cancellation, booking reference numbers and timestamps, and instructions for follow-up actions if needed. Input **variations** include: 1/ different accents, speech patterns, and dialects; 2/ background noise from various environments (airport, car, public space); 3/ varying levels of guest emotion or frustration; 4/ interruptions in the conversation flow; 5/ different ways of expressing the same confirmation or cancellation intent; and 6/ diverse speech clarity (mumbling, speaking too fast, etc.). The **error types** , ranked in order of estimated negative impact on stakeholders, include: 1/ failing to correctly identify the reservation, leading to cancellation of the wrong booking; 2/ failing to complete the cancellation in the hotel system while confirming to the guest; 3/ incorrectly calculating or communicating cancellation fees; 4/ misunderstanding critical guest information, resulting in unsuccessful cancellation; 5/ misinterpreting guest intent when they're asking questions vs. confirming actions; 6/ not recognizing guest speech due to accent or background noise; 7/ losing context during multi-turn conversations about cancellation details; and 8/ providing generic responses that don't address the guest's specific situation. With this in mind, we would expect the hotel customer service team to test an example prompt in the Console playground or API and review the output. +Consider the following use case of utilizing Amazon Nova 2 Sonic to power an AI agent that helps a hotel guest cancel their upcoming reservation through natural voice conversation via a phone call. The **business goal** is to provide an efficient, 24/7 self-service voice assistant that can handle hotel reservation cancellations while maintaining customer satisfaction and reducing call center volume. The **stakeholders** include the hotel guest, who wants to quickly cancel their reservation and receive confirmation without friction, and the hotel customer service team, who wants to automate routine cancellations to focus on more complex guest issues. The **workflow** is: 1/ the guest initiates a voice conversation with the customer-built hotel assistant application; 2/ the assistant application calls Amazon Nova 2 Sonic to generate speech output, requesting reservation details from the guest; 3/ the guest provides the necessary information through speech; 4/ the system verifies the reservation in the hotel's database; 5/ the system informs the guest of any applicable cancellation policies or fees; 6/ the guest confirms their desire to proceed with the cancellation; 7/ the system processes the cancellation in the hotel's reservation system; 8/ the system provides verbal confirmation; and 9/ the system sends a follow-up email receipt. In this workflow, Amazon Nova 2 Sonic handles speech understanding, generation, and dialog dynamics (steps 2, 3, 5, 6, and 8), while the customer-built assistant application manages the tool calling, database access, and business logic (steps 1, 4, 7, and 9). **Input prompts** contain information regarding reservation identification (confirmation number, booking dates), guest identification (name, contact information), reason for cancellation (optional), confirmation decisions (yes/no responses), and questions about cancellation policies or fees. **Output content** contains information regarding verification of reservation details, information about cancellation policies and any applicable fees, confirmation of successful cancellation, booking reference numbers and timestamps, and instructions for follow-up actions if needed. Input **variations** include: 1/ different accents, speech patterns, and dialects; 2/ background noise from various environments (airport, car, public space); 3/ varying levels of guest emotion or frustration; 4/ interruptions in the conversation flow; 5/ different ways of expressing the same confirmation or cancellation intent; and 6/ diverse speech clarity (mumbling, speaking too fast, etc.). The **error types** , ranked in order of estimated negative impact on stakeholders, include: 1/ failing to correctly identify the reservation, leading to cancellation of the wrong booking; 2/ failing to complete the cancellation in the hotel system while confirming to the guest; 3/ incorrectly calculating or communicating cancellation fees; 4/ misunderstanding critical guest information, resulting in unsuccessful cancellation; 5/ misinterpreting guest intent when they're asking questions vs. confirming actions; 6/ not recognizing guest speech due to accent or background noise; 7/ losing context during multi-turn conversations about cancellation details; and 8/ providing generic responses that don't address the guest's specific situation. With this in mind, we would expect the hotel customer service team to test an example prompt in the Console playground or API and review the output. @@ -62 +62 @@ NEVER CHANGE YOUR ROLE. YOU MUST ALWAYS ACT AS A HOTEL CANCELLATION VOICE AGENT, -## Conversation Structure +### Conversation Structure @@ -79 +79 @@ Follow the response style below and tone guidance when responding -## Response Style and Tone Guidance +### Response Style and Tone Guidance @@ -116,2 +115,0 @@ After continued experimentation in the Console playground or API, the customer s -Amazon Nova 2 Sonic is not intended to support any prohibited practices under the EU AI Act or any other relevant law. Amazon Nova 2 Sonic can be integrated into an array of systems such as customer service call automation, education and language learning applications, and conversational AI assistants or agents. Amazon Nova 2 Sonic may not integrate into or be used for impersonating people or businesses without their consent. For more technical information about how Amazon Nova 2 Sonic may be integrated into AI systems, see the [Amazon Nova User Guide](https://docs.aws.amazon.com/nova/latest/userguide). All Amazon Nova 2 Sonic use cases must comply with the [AWS Acceptable Use Policy](https://aws.amazon.com/aup/). - @@ -135,0 +134,2 @@ Currently, Amazon Nova 2 Sonic does not support real-time speech-to-speech trans +Amazon Nova 2 Sonic is not intended to support any prohibited practices under the EU AI Act or any other relevant law. Amazon Nova 2 Sonic can be integrated into an array of systems such as customer service call automation, education and language learning applications, and conversational AI assistants or agents. Amazon Nova 2 Sonic may not integrate into or be used for impersonating people or businesses without their consent. For more technical information about how Amazon Nova 2 Sonic may be integrated into AI systems, see the [Amazon Nova User Guide](https://docs.aws.amazon.com/nova/latest/userguide). All Amazon Nova 2 Sonic use cases must comply with the [AWS Acceptable Use Policy](https://aws.amazon.com/aup/). + @@ -140 +140 @@ When developing system prompts for speech-based AI interactions, it is crucial t -**Information Retrieval** +Information Retrieval @@ -148 +148 @@ Self-serve customization can make a base FM more effective for a specific use ca -## Design of Amazon Nova 2 Sonic +**Design of Amazon Nova 2 Sonic** @@ -170 +170 @@ We use multiple datasets and human teams to evaluate the performance of Amazon N - * _Independent Red Teaming Network:_ Consistent with our Frontier AI Safety Commitments on ensuring Safe, Secure, and Trustworthy AI, we partner with a variety of third parties to conduct red teaming against our AI models. We leverage red teaming firms to complement our in-house testing in areas such as safety, security, privacy, fairness, and veracity-related topics. We also work with specialized firms and academics to red-team our models for specialized areas such as Cybersecurity and Chemical, Biological, Radiological, and Nuclear (CBRN) capabilities. + * Independent Red Teaming Network: Consistent with our Frontier AI Safety Commitments on ensuring Safe, Secure, and Trustworthy AI, we partner with a variety of third parties to conduct red teaming against our AI models. We leverage red teaming firms to complement our in-house testing in areas such as safety, security, privacy, fairness, and veracity-related topics. We also work with specialized firms and academics to red-team our models for specialized areas such as Cybersecurity and Chemical, Biological, Radiological, and Nuclear (CBRN) capabilities. @@ -175 +175 @@ We use multiple datasets and human teams to evaluate the performance of Amazon N -### Safety +**Safety** @@ -185 +185 @@ We evaluate Amazon Nova 2 Sonic's capability to reject potentially harmful promp -Toxicity is a common, but narrow form of harmfulness, on which individual opinion varies widely. We assess Amazon Nova Lite 2.0's ability to avoid responding with content that contains potentially toxic content through automated testing on multiple datasets, and find that it performs well on common toxicity types. For example, on a proprietary toxic prompts dataset containing 8.5K prompts which we classified into sub-categories (for example, violence and gore, insults and stereotype, hate symbols, sexual content), Amazon Nova 2 Sonic's end-to-end guardrails provide safe responses to over 95% of the prompts. +Toxicity is a common, but narrow form of harmfulness, on which individual opinion varies widely. We assess Amazon Nova 2 Sonic's ability to avoid responding with content that contains potentially toxic content through automated testing on multiple datasets, and find that it performs well on common toxicity types. For example, on a proprietary toxic text prompts dataset containing 8.5K prompts which we classified into sub-categories (for example, violence and gore, insults and stereotype, hate symbols, sexual content), Amazon Nova 2 Sonic's end-to-end guardrails provide safe responses to over 95% of the prompts. @@ -193 +193 @@ To help prevent potential misuse, Amazon Bedrock implements automated abuse dete -Compared to information available via internet searches, science articles, and paid experts, we see no indications that Amazon Nova 2 Sonic increases access to information about, chemical, biological, radiological or nuclear (CBRN) threats. We continue to assess for CBRN risk, and engage with other third party researchers or vendors to share, learn about, and mitigate possible CBRN threats and vulnerabilities. +Compared to information available via internet searches, science articles, and paid experts, we see no indications that Amazon Nova 2 Sonic increases access to information about, chemical, biological, radiological or nuclear (CBRN) threats. We continue to assess for CBRN risk and engage with other third-party researchers or vendors to share, learn about, and mitigate possible CBRN threats and vulnerabilities. @@ -195 +195 @@ Compared to information available via internet searches, science articles, and p -### Fairness +**Fairness** @@ -199 +199 @@ Amazon Nova 2 Sonic is designed to conversationally interact with a diverse set -### Veracity +**Veracity** @@ -201 +201 @@ Amazon Nova 2 Sonic is designed to conversationally interact with a diverse set -Because transformer-based FMs are token generation engines, and not information retrieval engines, their outputs may contain statements that contradict statements in the prompt or that contradict facts verifiable from trusted third-party sources, or the outputs may omit statements that customers expect should be made, given information in the prompt or even just "common sense." Customers should carefully consider whether or not using RAG will improve the effectiveness of their solution; use of RAG can still result in errors. We assess Amazon Nova 2 Sonic's general knowledge without RAG on multiple datasets, and find that the models perform well, given the intrinsic limitations of large language models technology. +Because transformer-based FMs are token generation engines, and not information retrieval engines, their outputs may contain statements that contradict statements in the prompt or that contradict facts verifiable from trusted third-party sources, or the outputs may omit statements that customers expect should be made, given information in the prompt or even just "common sense." Customers should carefully consider whether or not using RAG will improve the effectiveness of their solution; use of RAG can still result in errors. We assess Amazon Nova 2 Sonic's general knowledge without RAG on multiple datasets and find that the model rarely generates inaccurate information and frequently include all expected facts. This indicates good veracity given the intrinsic limitations of large language models technology. @@ -203 +203 @@ Because transformer-based FMs are token generation engines, and not information -### Robustness +**Robustness** @@ -205 +205 @@ Because transformer-based FMs are token generation engines, and not information -We maximize robustness with a number of techniques, including using large training datasets that capture many kinds of variation across many different semantic intents. We measure model robustness by applying small, semantics-preserving perturbations to each and compare the responses to see how stable or invariant they are. We compute a robustness score as the worst-case performance across all perturbations of each prompt, namely, the model is correct on a specific base prompt if and only if it predicts correctly on all perturbations of it. +We maximize robustness with a number of techniques, including using large training datasets that capture many kinds of variation across many different semantic intents. We measure model robustness by applying small, semantics-preserving perturbations to each and compare the responses to see how stable or invariant they are. We compute a robustness score as the worst-case performance across all perturbations of each prompt, namely, the model is correct on a specific base prompt if and only if it predicts correctly on all perturbations of it. We find the model robust, with consistent correct prediction for perturbations of inputs including speech-specific perturbations and robustness to speech attributes. @@ -207 +207 @@ We maximize robustness with a number of techniques, including using large traini -### Explainability +**Explainability** @@ -211 +211 @@ Customers wanting to understand the steps taken by Amazon Nova 2 Sonic to arrive -### Privacy +**Privacy** @@ -213 +213 @@ Customers wanting to understand the steps taken by Amazon Nova 2 Sonic to arrive -Amazon Nova 2 Sonic is available in Amazon Bedrock. Amazon Bedrock is a managed service and does not store or review customer prompts or customer image generations, and prompts and generations are never shared between customers, or with Amazon Bedrock third party model providers. AWS does not use inputs or outputs generated through the Amazon Bedrock service to train Amazon Bedrock models, including Amazon Nova 2 Sonic. For more information, see Section 50.3 of the [AWS Service Terms](https://aws.amazon.com/service-terms/) and the [AWS Data Privacy FAQs](https://aws.amazon.com/compliance/data-privacy-faq/). For service-specific privacy information, see Security in the [Amazon Bedrock FAQs](https://aws.amazon.com/bedrock/faqs/). Amazon Nova models are designed to avoid completing prompts that could be construed as requesting private information. If a user is concerned that their private information has been included in an Amazon Nova model output, the user should contact us here. +Amazon Nova 2 Sonic is available in Amazon Bedrock. Amazon Bedrock is a managed service and does not store or review customer prompts or customer speech or text generations, and prompts and generations are never shared between customers, or with Amazon Bedrock third party model providers. During an active Amazon Nova 2 Sonic session and while speech is streamed continually, session context including conversation history is temporarily maintained in CPU/GPU cache and RAM as part of the memory hierarchy to enable real-time conversational flow. When the session ends or completes, request teardown causes all memory to be eligible for garbage collection following standard practices. No references are retained to objects from completed requests. AWS does not use inputs or outputs generated through the Amazon Bedrock service to train Amazon Bedrock models, including Amazon Nova 2 Sonic. For more information, see Section 50.3 of the [AWS Service Terms](https://aws.amazon.com/service-terms/) and the [AWS Data Privacy FAQs](https://aws.amazon.com/compliance/data-privacy-faq/). For service-specific privacy information, see Security in the [Amazon Bedrock FAQs](https://aws.amazon.com/bedrock/faqs/). Amazon Nova models are designed to avoid completing prompts that could be construed as requesting private information. If a user is concerned that their private information has been included in an Amazon Nova model output, the user should contact us [here](https://titan.aws.com/privacy). Amazon Nova 2 Sonic also includes guardrails designed to prevent voice cloning and impersonation. The model is designed to deflect prompts attempting to clone voices, helping protect individuals from unauthorized voice replication. @@ -215 +215 @@ Amazon Nova 2 Sonic is available in Amazon Bedrock. Amazon Bedrock is a managed -### Security +**Security** @@ -217 +217 @@ Amazon Nova 2 Sonic is available in Amazon Bedrock. Amazon Bedrock is a managed -All Amazon Bedrock models, including Amazon Nova 2 Sonic, come with enterprise security that enables customers to build generative AI applications that support common data security and compliance standards, including GDPR and HIPAA. Customers can use AWS PrivateLink to establish private connectivity between customized Amazon Nova 2 Sonic models and on-premises networks without exposing customer traffic to the internet. Customer data is always encrypted in transit and at rest, and customers can use their own keys to encrypt the data, for example, using AWS Key Management Service (AWS KMS). Customers can use IAM to securely control access to Amazon Bedrock resources. Also, Amazon Bedrock offers comprehensive monitoring and logging capabilities that can support customer governance and audit requirements. For example, CloudWatch can help track usage metrics that are required for audit purposes, and CloudTrail can help monitor API activity and troubleshoot issues as Amazon Nova 2 Sonic is integrated with other AWS systems. Customers can also choose to store the metadata, prompts, and image generations in their own encrypted Amazon S3 bucket. For more information, see [Amazon Bedrock Security](https://docs.aws.amazon.com/bedrock/latest/userguide/security.html). +All Amazon Bedrock models, including Amazon Nova 2 Sonic, come with enterprise security that enables customers to build generative AI applications that support common data security and compliance standards, including GDPR and HIPAA. Customers can use AWS PrivateLink to establish private connectivity between customized Amazon Nova 2 Sonic models and on-premises networks without exposing customer traffic to the internet. Customer data is always encrypted in transit and at rest, and customers can use their own keys to encrypt the data, for example, using AWS Key Management Service (AWS KMS). Customers can use IAM to securely control access to Amazon Bedrock resources. Also, Amazon Bedrock offers comprehensive monitoring and logging capabilities that can support customer governance and audit requirements. For example, CloudWatch can help track usage metrics that are required for audit purposes, and CloudTrail can help monitor API activity and troubleshoot issues as Amazon Nova 2 Sonic is integrated with other AWS systems. Customers can also choose to store the metadata, prompts, and speech or text generations in their own encrypted Amazon S3 bucket. For more information, see [Amazon Bedrock Security](https://docs.aws.amazon.com/bedrock/latest/userguide/security.html). @@ -219 +219 @@ All Amazon Bedrock models, including Amazon Nova 2 Sonic, come with enterprise s -### Intellectual Property +**Intellectual Property** @@ -221 +221 @@ All Amazon Bedrock models, including Amazon Nova 2 Sonic, come with enterprise s -Amazon Nova 2 Sonic is designed for conversational use-cases. We use guardrails to prevent customers from using our services to violate the rights of others, including both textual content and voice characteristics of speech. AWS offers uncapped intellectual property (IP) indemnity coverage for outputs of generally available Amazon Nova models (see Section 50.10 of the [AWS Service Terms](https://aws.amazon.com/service-terms/)). This means that customers are protected from third-party claims alleging IP infringement or misappropriation (including copyright claims) by the outputs generated by these Amazon Nova models. In addition, our standard IP indemnity for use of the Services protects customers from third-party claims alleging IP infringement (including copyright claims) by the Services (including Amazon Nova models) and the data used to train them. +Amazon Nova 2 Sonic is designed for conversational use-cases. We use guardrails to prevent customers from using our services to violate the rights of others, including both textual content and voice characteristics of speech. AWS offers uncapped intellectual property (IP) indemnity coverage for outputs of generally available Amazon Nova models (see Section 50.10 of the [AWS Service Terms](https://aws.amazon.com/service-terms/)). This means that customers are protected from third-party claims alleging IP infringement or misappropriation (including copyright claims) by the outputs generated by these Amazon Nova models. In addition, our standard IP indemnity for use of the Services protects customers from third-party claims alleging IP infringement (including copyright claims) by the Services (including Amazon Nova models) and the data used to train them. If you are a rightsholder/authorized representative and have a complaint regarding our commitments under the Copyright Chapter of the Code of Practice for General-Purpose AI Models under the EU AI Act, you may contact us at [[email protected]](mailto:[email protected]). Please be sure to include enough detail for us to investigate your complaint. @@ -223 +223 @@ Amazon Nova 2 Sonic is designed for conversational use-cases. We use guardrails -### Transparency +**Transparency** @@ -227 +227 @@ Amazon Nova 2 Sonic provides information to customers in the following locations -**Watermarking:** Amazon Nova 2 Sonic applies an inaudible watermark to all audio generated, helping identify AI-generated speech to promote the safe, secure, and transparent development of AI technology and helping reduce the spread of disinformation. If you have a specific request for detecting watermark for a given audio, please contact your AWS representative to discuss your needs. +Watermarking: Amazon Nova 2 Sonic applies an inaudible watermark to all audio generated, helping identify AI-generated speech to promote the safe, secure, and transparent development of AI technology and helping reduce the spread of disinformation. If you have a specific request for detecting watermark for a given audio, please contact your AWS representative to discuss your needs. @@ -229 +229 @@ Amazon Nova 2 Sonic provides information to customers in the following locations -### Governance +**Governance** @@ -233 +233 @@ We have rigorous methodologies to build our AWS AI services responsibly, includi -## Deployment and performance optimization best practices +**Deployment and performance optimization best practices** @@ -256 +256 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -## Further information +**Further information** @@ -258 +258 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi - * For service documentation, see [Amazon Nova 2 Sonic User Guide](https://docs.aws.amazon.com/nova/latest/userguide). + * For service documentation, see [Amazon Nova 2 Sonic User Guide](https://docs.aws.amazon.com/nova/latest/nova2-userguide/what-is-nova-2.html). @@ -260 +260 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi - * For details on privacy and other legal considerations, see AWS's [Acceptable Use Policy](https://aws.amazon.com/aup/), [Responsible AI Policy](https://aws.amazon.com/ai/responsible-ai/policy/), [Legal](https://aws.amazon.com/legal/), [Compliance](https://aws.amazon.com/compliance/), [Privacy](https://aws.amazon.com/privacy/). + * For details on privacy and other legal considerations, see AWS’s [Acceptable Use Policy](https://aws.amazon.com/aup/), [Responsible AI Policy](https://aws.amazon.com/machine-learning/responsible-ai/policy/), [Legal](https://aws.amazon.com/legal/), [Compliance](https://aws.amazon.com/compliance/), [Privacy](https://aws.amazon.com/privacy/). @@ -262 +262 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi - * For help optimizing a workflow, see [Generative AI Innovation Center](https://aws.amazon.com/ai/generative-ai/innovation-center/), [AWS Customer Support](https://aws.amazon.com/contact-us/), [AWS Professional Services](https://aws.amazon.com/professional-services/), [Amazon SageMaker Ground Truth Plus](https://aws.amazon.com/sagemaker/groundtruth/), [AWS Well-Architected](https://aws.amazon.com/wellarchitected/). + * For help optimizing a workflow, see [Generative AI Innovation Center](https://aws.amazon.com/generative-ai/innovation-center/),[AWS Customer Support](https://aws.amazon.com/contact-us/), [AWS Professional Services](https://aws.amazon.com/professional-services/), [Amazon SageMaker AI Ground Truth Plus](https://aws.amazon.com/sagemaker/data-labeling/),[AWS Well-Architected.](https://aws.amazon.com/premiumsupport/business-support-well-architected/?trk=64889dc8-355d-4be1-8abb-b78d97ad2a1f&sc_channel=ps&ef_id=Cj0KCQjwgvnCBhCqARIsADBLZoLG-He86LFiiJ6Z9mez3_JAdEGg-wL0qKI5kxQjdKTUkr9JygW8ugUaAiAGEALw_wcB:G:s&s_kwcid=AL!4422!3!719333527370!e!!g!!aws%20well%20architected!21852254337!176452274104&gad_campaignid=21852254337&gclid=Cj0KCQjwgvnCBhCqARIsADBLZoLG-He86LFiiJ6Z9mez3_JAdEGg-wL0qKI5kxQjdKTUkr9JygW8ugUaAiAGEALw_wcB) @@ -264 +264 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi - * For other tools to help customers work with foundation models, see [Amazon Bedrock](https://docs.aws.amazon.com/bedrock/), [Amazon Bedrock Guardrails](https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html), [Amazon Bedrock Guardrails automated reasoning checks](https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails-automated-reasoning-checks.html), [Amazon Q developer](https://aws.amazon.com/q/developer/), and Nova Understanding Models. + * For other tools to help customers work with foundation models, see [Amazon Bedrock](https://aws.amazon.com/bedrock/?trk=0eaabb80-ee46-4e73-94ae-368ffb759b62&sc_channel=ps&ef_id=Cj0KCQjwmK_CBhCEARIsAMKwcD7_qJb4tLoHPg3oOKsFq_dCSaH8VHgAK2B7XI0q-FHuXkVwFaumhjAaAq4IEALw_wcB:G:s&s_kwcid=AL!4422!3!692006004688!p!!g!!amazon%20bedrock!21048268554!159639952935&gad_campaignid=21048268554&gclid=Cj0KCQjwmK_CBhCEARIsAMKwcD7_qJb4tLoHPg3oOKsFq_dCSaH8VHgAK2B7XI0q-FHuXkVwFaumhjAaAq4IEALw_wcB), [Amazon Bedrock Guardrails](https://aws.amazon.com/bedrock/guardrails/), [Amazon Bedrock Guardrails automated reasoning checks](https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-bedrock-guardrails-automated-reasoning-checks-preview/), [Amazon Q developer](https://docs.aws.amazon.com/amazonq/latest/qdeveloper-ug/what-is.html), and [Nova Understanding Models.](https://nova.amazon.com/chat) @@ -271 +271,7 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -## Glossary +**Glossary** + + * **Controllability:** Steering system behavior to reflect system design goals + + * **Privacy & Security**: Appropriately obtaining, using and protecting data and models + + * **Safety:** Preventing harmful system output and misuse @@ -273 +279 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Controllability:** Steering system behavior to reflect system design goals + * **Fairness:** Considering impacts on different groups of stakeholders @@ -275 +281 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Privacy & Security: **Appropriately obtaining, using and protecting data and models + * **Explainability:** Understanding and evaluating system outputs @@ -277 +283 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Safety:** Preventing harmful system output and misuse + * **Veracity & Robustness:** Achieving correct system outputs, even with unexpected or adversarial inputs @@ -279 +285 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Fairness:** Considering impacts on different groups of stakeholders + * **Transparency:** Enabling stakeholders to make informed choices about their engagement with an AI system @@ -281 +287 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Explainability:** Understanding and evaluating system outputs + * **Governance:** Embedding best practices within the AI supply chain, including providers and deployers @@ -283 +288,0 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Veracity & Robustness: **Achieving correct system outputs, even with unexpected or adversarial inputs @@ -285 +289,0 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Transparency:** Enabling stakeholders to make informed choices about their engagement with an AI system @@ -287 +290,0 @@ The performance of any application using Amazon Nova 2 Sonic depends on the desi -**Governance:** Embedding best practices within the AI supply chain, including providers and deployers