AWS ai documentation change
Summary
Formatting adjustments to example conversation and grammatical changes (e.g., 'customers' to 'you', 'their' to 'your') to improve readability and direct address. No technical content changes.
Security assessment
Changes are grammatical or formatting-focused. Security-related content (e.g., compliance, safeguards, responsible AI practices) was already present and remains unchanged in substance. No evidence of addressing vulnerabilities or new security features.
Diff
diff --git a/ai/responsible-ai/nova-sonic/overview.md b/ai/responsible-ai/nova-sonic/overview.md index 4424d2c24..46d5dfe89 100644 --- a//ai/responsible-ai/nova-sonic/overview.md +++ b//ai/responsible-ai/nova-sonic/overview.md @@ -84 +84,7 @@ Here is an example of a sample prompt and the subsequent voice conversation with -[User] Please proceed with the cancellation. [Amazon Nova Sonic] Great, I'll proceed with the cancellation for you. Let me confirm that your reservation has been successfully cancelled. If you have any further questions or need additional assistance, please don't hesitate to contact us. Have a wonderful day! [User] Thank you, bye!Assessing the completion for effectiveness, we observe a/ no contradictions of the facts in the prompt, b/ no toxic or unsafe speech, c/ key product information present, and d/ coherent and organized response. After continued experimentation with the prompt, the customer should finalize their own measure of effectiveness based on the impact of errors, run a scaled-up test via the API, and use the results of human judgements (with multiple judgements per test prompt) to establish a benchmark effectiveness score. Amazon Bedrock directly offers these kinds of testing capabilities. +[User] Please proceed with the cancellation. + +[Amazon Nova Sonic] Great, I'll proceed with the cancellation for you. Let me confirm that your reservation has been successfully cancelled. If you have any further questions or need additional assistance, please don't hesitate to contact us. Have a wonderful day! + +[User] Thank you, bye! + +Assessing the completion for effectiveness, we observe a/ no contradictions of the facts in the prompt, b/ no toxic or unsafe speech, c/ key product information present, and d/ coherent and organized response. After continued experimentation with the prompt, you should finalize your own measure of effectiveness based on the impact of errors, run a scaled-up test via the API, and use the results of human judgements (with multiple judgements per test prompt) to establish a benchmark effectiveness score. Amazon Bedrock directly offers these kinds of testing capabilities. @@ -91 +97 @@ Amazon Nova Sonic has a number of limitations that require careful consideration -We make every effort to design, develop, and rigorously test our models to help ensure they produce appropriate outputs based on user inputs, but foundation models are by their nature non-deterministic and may occasionally produce unintended or undesirable outputs. We encourage users to provide feedback [here](https://pages.awscloud.com/global-ln-gc-400-ai-service-cards-contact-us-registration.html) about our models to help us continuously improve their performance. Customers should evaluate outputs for accuracy and appropriateness for their use cases, especially if these will be directly surfaced to end-users. Additionally, if Amazon Nova Sonic model is used in customer workflows that produce consequential decisions, customers must evaluate the potential risks of their use case and implement appropriate human oversight, testing, and other use case-specific safeguards to mitigate such risks. See the [AWS Responsible AI Policy](https://aws.amazon.com/ai/responsible-ai/policy/) for more information. Customers who use Amazon Nova Sonic are responsible for ensuring that their use of Amazon Nova Sonic and the generated speech or other output complies with all applicable laws. Amazon Nova Sonic and output may not be used for any prohibited practices under the EU AI Act. +We make every effort to design, develop, and rigorously test our models to help ensure they produce appropriate outputs based on user inputs, but foundation models are by their nature non-deterministic and may occasionally produce unintended or undesirable outputs. We encourage users to provide feedback [here](https://pages.awscloud.com/global-ln-gc-400-ai-service-cards-contact-us-registration.html) about our models to help us continuously improve their performance. You should should evaluate outputs for accuracy and appropriateness for their use cases, especially if these will be directly surfaced to end-users. Additionally, if Amazon Nova Sonic model is used in your workflows that produce consequential decisions, you must evaluate the potential risks of their use case and implement appropriate human oversight, testing, and other use case-specific safeguards to mitigate such risks. See the [AWS Responsible AI Policy](https://aws.amazon.com/ai/responsible-ai/policy/) for more information. If you use Amazon Nova Sonic you are responsible for ensuring that your use of Amazon Nova Sonic and the generated speech or other output complies with all applicable laws. Amazon Nova Sonic and output may not be used for any prohibited practices under the EU AI Act. @@ -96 +102 @@ We make every effort to design, develop, and rigorously test our models to help -Amazon Nova Sonic is designed to disengage with attempts to circumvent its safety measures through prompt engineering. If a customer's speech generation request is unsuccessful, it may be due to one or more such measures. The safety filters for Amazon Nova Sonic cannot be configured or turned off. However, they are periodically assessed and improved in response to feedback. +Amazon Nova Sonic is designed to disengage with attempts to circumvent its safety measures through prompt engineering. If a speech generation request is unsuccessful, it may be due to one or more such measures. The safety filters for Amazon Nova Sonic cannot be configured or turned off. However, they are periodically assessed and improved in response to feedback. @@ -116 +122 @@ When developing system prompts for speech-based AI interactions, it's crucial to -By itself, Amazon Nova Sonic is not an information retrieval tool. The Amazon Nova Sonic model training corpus does not cover all dialects, cultures, geographies and time periods, or the domain specific knowledge a customer may need for a particular use case. We do not define a "cutoff date" for training or otherwise try to characterize the foundation model as a knowledge base. Customers with workflows requiring accurate information from a specific knowledge domain or time period should consider employing tool use for knowledge grounding. +By itself, Amazon Nova Sonic is not an information retrieval tool. The Amazon Nova Sonic model training corpus does not cover all dialects, cultures, geographies and time periods, or the domain specific knowledge you may need for a particular use case. We do not define a "cutoff date" for training or otherwise try to characterize the foundation model as a knowledge base. If you have workflows requiring accurate information from a specific knowledge domain or time period, you should consider employing tool use for knowledge grounding. @@ -121 +127 @@ By itself, Amazon Nova Sonic is not an information retrieval tool. The Amazon No -Self-serve customization can make a base FM more effective for a specific use case, particularly for more compact models that offer lower cost. However, customers cannot fine-tune Amazon Nova Sonic on their own labeled data; they can only customize it using system prompts. We will add self-service fine-tuning capabilities to Amazon Nova Sonic in near future. For more information, see the [Amazon Nova Sonic User Guide](https://docs.aws.amazon.com/nova/latest/userguide/speech.html). +Self-serve customization can make a base FM more effective for a specific use case, particularly for more compact models that offer lower cost. However, you can't fine-tune Amazon Nova Sonic on your own labeled data; you can only customize it using system prompts. We will add self-service fine-tuning capabilities to Amazon Nova Sonic in near future. For more information, see the [Amazon Nova Sonic User Guide](https://docs.aws.amazon.com/nova/latest/userguide/speech.html). @@ -137 +143 @@ In both scenarios, the user prompts the Amazon Nova Sonic model with a system pr -In application A, Amazon Nova Sonic must handle background call center noise, interpret domain-specific terminology, and retrieve accurate information from their knowledge base. For application B, Amazon Nova Sonic needs to support instructional activities across multiple subjects. Amazon Nova Sonic must process subject-specific terminology, maintain context throughout instructional sequences and adapt to different pedagogical approaches. Environmental factors (background noise, device quality), linguistic variations (dialect, accent), integration complexity (custom knowledge bases, authentication systems), conversation patterns (multi-turn complexity, context retention), and deployment constraints (latency requirements, failover mechanisms) all influence real-world performance. Because performance results depend on a variety of factors including the Amazon Nova Sonic, the customer workflow, and the evaluation dataset, we recommend that customers test the model using their own content. Amazon Bedrock and Amazon SageMaker AI Clarify directly provide automated and human testing capabilities. +In application A, Amazon Nova Sonic must handle background call center noise, interpret domain-specific terminology, and retrieve accurate information from their knowledge base. For application B, Amazon Nova Sonic needs to support instructional activities across multiple subjects. Amazon Nova Sonic must process subject-specific terminology, maintain context throughout instructional sequences and adapt to different pedagogical approaches. Environmental factors (background noise, device quality), linguistic variations (dialect, accent), integration complexity (custom knowledge bases, authentication systems), conversation patterns (multi-turn complexity, context retention), and deployment constraints (latency requirements, failover mechanisms) all influence real-world performance. Because performance results depend on a variety of factors including the Amazon Nova Sonic, the customer workflow, and the evaluation dataset, we recommend that you test the model using their own content. Amazon Bedrock and Amazon SageMaker AI Clarify directly provide automated and human testing capabilities.