Privacy Controls and Technology Direction
SENEX describes privacy technology by the protection it should provide, not by publishing a recipe for the underlying implementation. Technology names alone do not establish privacy; controls must be connected to a specific data flow, threat, and body of evidence.
The current foundation is local-first AIA development: keep user-selected context in the user-controlled environment, require deliberate action for external exchange, and design sharing around a bounded purpose. Coverage and behavior remain under active development.
Control layers
Local data control
The first layer limits unnecessary collection. AIA should let users choose which sources can be used, understand when context is active, and remove access that is no longer wanted. Sensitive processing should remain local where the task permits it.
Local processing does not eliminate device risk. Compromised accounts, unsafe backups, other software, physical access, and operating-system behavior can all cross the same boundary. Product guidance and testing must account for this wider environment.
Purpose-bound exchange
When information must leave the local environment, the intended control is a narrowly scoped request and a bounded response. The user should be able to distinguish a one-time action from continuing access and understand the recipient and purpose.
Protection after disclosure has limits. Revocation can stop future authorized access, but it cannot erase knowledge already received by an independent party. Minimization before transfer is therefore more important than relying on revocation alone.
Integrity and accountability
Future cooperation requires evidence that a request came from an authorized participant, that the approved content was not silently changed in transit, and that a relevant event can be investigated. Public documentation states these outcomes without exposing internal identity, messaging, audit, or validation constructions.
Authenticated network cooperation, consistent event records, and contribution-aware accountability are testnet targets. V1-testnet will use test-only value and is not a production privacy network.
Privacy-enhancing collaboration
SENEX is researching ways for participants to cooperate while reducing what any participant can learn beyond the intended result. Candidate approaches will be selected and evaluated privately against defined threats, performance constraints, and failure behavior.
No formal privacy-preserving collaboration mechanism is claimed as a current feature. Protocol choices, parameters, and internal evaluation details are intentionally not published before implementation and review.
How a control becomes a claim
A privacy control should move through an evidence path:
- define the protected information and the trust boundary;
- state the threat and expected failure behavior;
- implement the control without widening unrelated access;
- test normal, adversarial, and recovery scenarios;
- document limitations in user-facing language; and
- obtain independent review before making a strong public claim.
Implementation review, adversarial testing, and deployment-specific assessment are necessary before any control is described as formally private, anonymous, tamper-proof, or suitable for regulated use.
See Privacy Assurance, Threat Model, and Compliance for the evidence and claim boundaries.