Thirst AI Data Protection

Depending on the user’s permissions and the task being performed, Thirst AI may process:

• user prompts and relevant conversation context;
• profile, learning, skills and assignment information;
• course and content information;
• reporting information;
• authorised team information;
• uploaded files and other user-provided content.

Thirst AI does not give users access to information they would not otherwise be authorised to access through the platform. Customers can also control whether Thirst AI is enabled and which users or roles can access it.

Yes.

Thirst typically holds limited structured personal data, such as name, email address and job title. Additional personal information, including special-category data, may appear incidentally in free text, prompts or uploaded content.

Users should avoid including unnecessary personal or sensitive information in AI prompts.

Thirst AI uses Amazon Web Services and Amazon Bedrock.

Generative AI inference is restricted to the AWS London region (eu-west-2). Amazon Titan Text Embeddings V2, used to support AI memory functionality, is also configured in London.

Thirst uses Anthropic Claude through Amazon Bedrock rather than Anthropic’s direct API.

Amazon Web Services provides the cloud and AI infrastructure used by Thirst AI through Amazon Bedrock. Anthropic provides the Claude foundation model, but does not receive or have access to Thirst customer prompts or responses through Amazon Bedrock.

Thirst does not use customer prompts or AI outputs to train AI models. Amazon Bedrock does not use customer prompts or outputs to train foundation models.

Temporary AI-related data is retained as follows:

• conversation transcripts: 1 day;
• user/session IDs: approximately 25 hours;
• temporary tokens: approximately 60 seconds;
• AI-generated exports and intermediate query datasets: 7 days;
• CloudWatch logs: 30 days;
• X-Ray tracing: 30 days.

Where AI-generated content is deliberately saved into Thirst, it follows the normal content and customer retention arrangements.

Thirst AI can:

• answer questions and recommend relevant learning;
• provide learning and skills information;
• support reporting, exports and administrative tasks;
• generate, improve, summarise and transform learning content.

AI-generated content is provided as draft content for review before publication. Learning recommendations are advisory.

Thirst AI does not autonomously determine:

• assessment results or pass/fail outcomes;
• qualifications;
• access to learning;
• learner progression;
• other legally or similarly significant decisions about users.

Personalised recommendations may use learning-related information such as skills, assignments and activity, but remain advisory.

Amazon Web Services is the cloud and AI infrastructure provider used for Thirst AI and processes AI data through Amazon Bedrock.

Anthropic provides the Claude foundation model used through Amazon Bedrock, but does not receive or have access to Thirst customer prompts or responses.

For more information about how Thirst processes personal data and your data protection rights, please see our Privacy Notice.