Songwriter Bot Technical Specifications

The Your Voice is Power Songwriter Bot is meant to be used with the Your Voice is Power "What is the Role of AI in Art?" lesson.

The lesson explains AI concepts, explores AI best practices and limitations, explores legal and ethical considerations such as plagiarism and copyright, and drives discussions around the role of generative AI in creative processes such as songwriting.

This page contains technical details describing the technology used, privacy policies, and moderation features in place to protect students using the Songwriter Bot.

This page also contains links to details on sustainability commitments from the Large Language Model provider and the hosting provider.

Accounts required

• Students do not need to create accounts to use this service
• Teachers do not need to create accounts to use this service

Technologies used

Large Language Model

The Songwriter Bot uses the model Anthropic Claude 3.5 Sonnet 20240620 version 1.0

Detailed specifications can be found on the Claude 3 Model Card.

Cloud Service

The Songwriter Bot uses the service Amazon Bedrock to access the Anthropic Claude model.

Data Privacy

Large Language Model

Section 2.5 Training Data of the Claude 3 Model Card states the following:

"Claude 3 models are trained on a proprietary mix of publicly available information on the Internet as of August 2023, as well as non-public data from third parties, data provided by data labeling services and paid contractors, and data we generate internally. We employ several data cleaning and filtering methods, including deduplication and classification. The Claude 3 suite of models have not been trained on any user prompt or output data submitted to us by users or customers, including free users, Claude Pro users, and API customers. When Anthropic obtains data by crawling public web pages, we follow industry practices with respect to robots.txt instructions and other signals that website operators use to indicate whether they permit crawling of the content on their sites. In accordance with our policies, Anthropic’s crawler does not access password protected or sign-in pages or bypass CAPTCHA controls, and we conduct diligence on the data that we use. Anthropic operates its crawling system transparently, which means website operators can easily identify Anthropic visits and signal their preferences to Anthropic."

Hosting Service

As stated in the Amazon Bedrock Data Protection policy, "Amazon Bedrock doesn't store or log your prompts and completions. Amazon Bedrock doesn't use your prompts and completions to train any AWS models and doesn't distribute them to third parties".

The Songwriter Bot has implemented Amazon Bedrock sensitive information filters to remove Personally Identifiable Information (PIIs) from input prompts and model responses. The filters block the following PII:

• EMAIL
• PHONE
• USERNAME
• PASSWORD
• LICENSE_PLATE
• CREDIT_DEBIT_CARD_CVV
• VEHICLE_IDENTIFICATION_NUMBER
• PIN
• US_SOCIAL_SECURITY_NUMBER
• US_BANK_ACCOUNT_NUMBER
• INTERNATIONAL_BANK_ACCOUNT_NUMBER
• SWIFT_CODE, IP_ADDRESS, MAC_ADDRESS
• AWS_ACCESS_KEY
• AWS_SECRET_KEY
• CA_HEALTH_NUMBER
• CA_SOCIAL_INSURANCE_NUMBER
• CREDIT_DEBIT_CARD_NUMBER

The filters do not block the following PII but the workshop does explain to the students they should not share any private information with generative AI tools :

• ADDRESS - allowing ADDRESS allows the student to ask the Songwriter Bot to create lyrics that reflect a particular town or community. If ADDRESS PII is blocked students, cannot specify a town or Indigenous community in their prompt for song lyrics.
• NAME - allowing NAME allows the student to ask the Songwriter Bot to create lyrics about the role model stories provided in the lesson without having to remove their names from the provided text. If NAME PII is blocked, students cannot specify the name of a person in their prompt for song lyrics.

Protection from harmful content

Large Language Model

Section 2.6 Training process of the Claude 3 Model Card states the following:

"Claude was trained with a focus on being helpful, harmless, and honest. Training techniques include pre training on large diverse data to acquire language capabilities through methods like word prediction, as well as human feedback techniques that elicit helpful, harmless, honest responses. Anthropic used a technique called Constitutional AI [16] to align Claude with human values during reinforcement learning by explicitly specifying rules and principles based on sources like the UN Declaration of Human Rights. With Claude 3 models, we have added an additional principle to Claude’s constitution to encourage respect for disability rights, sourced from our research on Collective Constitutional AI [17]. Some of the human feedback data used to finetune Claude was made public [18] alongside our RLHF [19] and red-teaming research. Once our models are fully trained, we run a suite of evaluations for safety. Our Trust and Safety team also runs continuous classifiers to monitor prompts and outputs for harmful, malicious use cases that violate our AUP. See more on both in the evaluations sections."

Hosting Service

The Songwriter Bot uses Amazon Bedrock content filters to help detect and filter harmful user inputs and model-generated outputs. Below is a list of the filters applied and their settings.

Sustainability

Large Language Model

Section 4.3 Sustainability of the Claude 3 Model Card states the following:

"We offset our emissions (including from our cloud computing usage) and work with cloud providers that prioritize renewable energy and carbon neutrality. Anthropic works to fully offset our operational carbon emissions each year, partnering with external experts to conduct a rigorous analysis of our company-wide carbon footprint. Once measured, we invest in verified carbon credits to fully offset our annual footprint. Our credits directly fund emissions reduction projects. Our goal is to maintain net zero climate impact on an annual basis through such initiatives and offsets."

Hosting Service

AWS states "Our global infrastructure is designed and built to help make your workload more energy efficient, which is in line with our commitment to reach net-zero carbon by 2040." Refer to the AWS Sustainability page for additional details on their actions and commitment.