Understand. Decide. Apply.

Encryption: in transit and at rest

A confidential message needs more than a padlock in your browser. What matters is who can read it during transfer, in storage and while it is being processed.

AI LabsUpdated 3 sources

Two layers of protection

Browser

Readable input on an authorised device.

TLSIn transit

AI Link

Authorised services decrypt what they need to process.

AES-256-GCMAt rest

Stored secrets

Encrypted credentials with an authentication tag.

The diagram separates transport protection from application-level encryption of secrets. It does not imply that every stored document uses this same layer.

In transit: protecting the connection

TLS protects a connection between two endpoints, such as a browser and a server. Certificates help identify the server; encryption protects the content from interception during transfer. TLS 1.3 is specified in RFC 8446.

The receiving server can read the content. HTTPS alone therefore does not prevent an AI provider from processing a submitted text. Data minimisation and an appropriate model route provide additional protection.

At rest: AES-256-GCM for stored secrets

AI Link uses application-level AES-256-GCM for stored credentials and secrets, including access tokens for connected services. AES-256 uses a 256-bit key. GCM adds an authentication tag that allows alterations to the encrypted content to be detected.

This layer is distinct from encryption of entire disks or databases. It is also not end-to-end encryption against the processing service: authorised processes need access to carry out their work.

What your IT team should ask

A transparent security architecture describes the full scope of protection: encrypted data, key management and backup restoration. Together with provider selection and access controls, this supports an informed technical assessment.

  • Review transport, storage and processing separately
  • Limit key access to services that need it
  • Test recovery and key changes together

Sources & further reading

Original research, provider documentation and further reading on the concepts and figures discussed here.

Talk to AI LABS

Bring AI into the business. Securely and step by step.

A personal conversation clarifies suitable use cases, data protection requirements and the next steps, from a first demo to a pilot.

Get in touch