JamA11y AI Security Policy
Accessibility, AI, and Data Security
JamA11y uses AI to transform inaccessible PDFs into fully accessible documents - so that every student can engage with the content they need.
We don't sell or rent user data to any third parties.
Your documents are yours. JamA11y never monetises the content you upload or the data associated with your account.
No user data is used to train third-party or in-house AI models.
Documents uploaded for remediation are processed solely for the purpose of making them accessible. Neither JamA11y nor its AI partners use your content to improve, train, or develop any AI model.
Uploaded documents belong to the original content owner.
Neither JamA11y nor its users assume ownership of documents processed through the platform. Copyright remains with whoever held it before the document was uploaded. JamA11y exists to make your content accessible - not to claim it.
Basically, we do everything we can to ensure the ethical use of AI as a force for good in education. We help institutions make content accessible – we don't harvest their data.
Which third parties provide AI services to JamA11y?
Amazon Web Services (AWS) Bedrock
AWS provides the cloud infrastructure and AI services that power JamA11y's document processing pipeline. This includes optical character recognition (OCR), document structure analysis, alt text generation for images, and handling of complex content such as STEM equations and tables.
AWS does not use JamA11y user data to train its AI models.
Which third parties provide AI services to JamA11y?
Amazon Web Services (AWS) Bedrock
AWS provides the cloud infrastructure and AI services that power JamA11y's document processing pipeline. This includes optical character recognition (OCR), document structure analysis, alt text generation for images, and handling of complex content such as STEM equations and tables.
AWS does not use JamA11y user data to train its AI models.
To be clear, our relationship with AWS means that user data is:
Not used to train AWS AI models
Not used to train any third-party models
Not used to train any in-house JamA11y models
Not used to improve any third-party products or services
AWS cloud servers are located in regions close to the user. Data processed in the UK is hosted in the UK; data processed in the US is hosted in the US.
Who owns documents processed through JamA11y?
The simple answer: whoever owned the document before uploading it still owns it.
Processing a document through JamA11y does not transfer or affect ownership in any way. JamA11y honours the rights of original content creators - including publishers, faculty, and institutions - and does not allow uploaded source files to be redistributed through the platform.
Data segregation and account security
All user data in JamA11y is segregated and locked to individual institution accounts. No organisation can access another organisation's documents or account data.
JamA11y is SOC II compliant.
For any questions about JamA11y's data practices, please get in touch.