Glossary

Search results

Multi-Party Computation(also: MPC, Secure Multi-Party Computation, SMPC)

Multi-party computation (MPC) is a subfield of cryptography that enables multiple parties to jointly compute a function over their private inputs while keeping those inputs secret from each other. No single party learns anything beyond the output. In accessible digital systems,…

Obfuscation(also: Content Obfuscation, Visual Obfuscation)

AI-enabled techniques that automatically detect and remove or conceal private content in images and videos by applying filters such as blurring, masking, or blocking. In the context of visual assistance technologies, obfuscation aims to protect blind users from inadvertently…

Off-Device Processing(also: Cloud Processing, Remote Processing)

Computing performed on external servers rather than on the user's local device, requiring data to be transmitted over the internet. In the context of visual assistance technologies, off-device processing raises privacy and security concerns because users' images and videos —…

On-Device Processing(also: Local Processing, Edge Processing)

Computing performed directly on a user's device (phone, tablet, or wearable) rather than sending data to external servers. On-device processing is particularly valuable for privacy-sensitive accessibility features because it keeps personal visual data under the user's control,…

On-device Recognition(also: On-Device Inference, Edge Recognition)

Performing pattern recognition - such as sign language recognition, speech recognition, or computer vision - locally on the user's device rather than by sending input to a remote server. On-device recognition matters for accessibility because it preserves privacy (camera or…

Personally Identifiable Information(also: PII)

Any data that can be used to identify a specific individual, such as name, email address, location, biometric data, or device identifiers. For assistive technology users, PII concerns are heightened because the data collected often reveals sensitive information about a person's…

Personally Identifying Information(also: PII, Personal Data, Personally Identifiable Information)

Any data that can be used to identify a specific individual, including names, addresses, photographs, financial details, and biometric data. In accessibility contexts, PII is a significant concern when disabled users contribute data for AI training, as they may inadvertently…

Privacy

The right and practical ability of a person to control the collection, use, and disclosure of information about themselves, their body, their activities, and their relationships. For accessibility, privacy intersects with disability in specific ways: assistive-technology usage…

Privacy Alert(also: Privacy Notification, Privacy Warning)

A notification informing users about the potential presence of private or sensitive content in their captured images or video. Privacy alerts aim to prompt users to apply obfuscation or take other protective action. However, research with blind users highlights tensions: while…

Privacy Enhancing Technology(also: PET, PETs)

A category of tools, protocols, and practices designed to protect user privacy while using digital systems, including anonymising proxies, tracker blockers, end-to-end encryption, differential privacy, and private browsing modes. PETs increasingly intersect with accessibility…

Privacy Leak(also: Accidental Disclosure, Privacy Disclosure)

The unintentional capture and sharing of sensitive personal information through visual assistance technologies. Research has found that approximately 10% of images submitted to VAT services contain private content such as pregnancy tests, prescription medication, and people,…

Privacy Threat Model(also: Privacy Threat Analysis, Privacy Risk Assessment)

A systematic process for identifying, classifying, and evaluating potential privacy risks that a technology system may pose to its users. Privacy threat modeling extends security-focused frameworks (like Microsoft's STRIDE) to address privacy-specific concerns. The LINDDUN…

Privacy by Default(also: Privacy by Design, Default Privacy Settings)

Privacy by default is a design principle requiring that systems automatically protect user privacy without requiring users to take action. In accessibility contexts, this principle is particularly important for older adults and people with cognitive disabilities who may not…

Privacy by Design(also: PbD)

A framework that embeds privacy protections into the design and architecture of systems and business practices from the outset, rather than adding them as afterthoughts. In the context of visual assistance technologies, privacy by design encompasses on-device processing, data…

Privacy-Enhancing Data Filters(also: Privacy Filters, Data Obfuscation Filters)

Visual or data modifications applied to training datasets that obscure the identity of contributors while preserving the information needed for machine learning tasks. In the context of sign language video, these filters may include face blurring, cel shading, avatar…

Privacy-Enhancing Technologies(also: PETs)

Tools and techniques that protect user privacy by limiting the collection, use, and dissemination of personal information online. Examples include VPNs, ad blockers, privacy-focused browsers like Brave, encrypted messaging apps, and Do Not Track settings. For users with…

Private Visual Content(also: PVC, Visual Privacy)

Private visual content (PVC) refers to visual information in images or videos that the person depicted or sharing the content considers private and would not want publicly disclosed. For people who are blind using visual interpreter services, PVC is a particular concern because…

Pseudonymization(also: Pseudonymisation, De-identification)

A privacy technique in which personally identifying fields are replaced with artificial identifiers — typically hashes, tokens, or randomly assigned IDs — so that the data can no longer be attributed to a specific person without additional information kept separately. Recognised…

Push-to-Talk(also: PTT, Push to talk)

An interaction pattern where a user presses and holds (or taps) a dedicated button to signal the start of an input — historically used in two-way radios, now common in voice assistants and conversational interfaces as an alternative to continuous listening. In accessibility…

Re-identification risk(also: De-anonymization risk, Data re-identification)

The possibility that an individual can be identified from supposedly anonymized data by combining multiple data points or matching against external datasets. People with disabilities face heightened re-identification risk because uncommon combinations of attributes — rare…

Reverse Privacy Paradox

The reverse privacy paradox is a pattern, described by Zhang and colleagues in research on LLM-based conversational agents, in which users appear to disregard privacy concerns in the moment of use while still recognising those concerns exist and being willing to adopt…

Revocable Consent(also: Withdrawable Consent)

A consent pattern in which the user can withdraw their previously granted permission at any time, typically through a persistent, discoverable UI control that immediately halts data processing and triggers deletion of data collected under that consent. A stronger form than…

Right to Erasure(also: Right to be Forgotten, GDPR Article 17)

A user right under the EU General Data Protection Regulation (Article 17) to have their personal data deleted by a data controller when certain conditions are met (e.g., data no longer needed, consent withdrawn, unlawful processing). Implemented in accessible products through…

Role-Based Access Control(also: RBAC)

An authorisation model in which system permissions are attached to roles (e.g., user, administrator, clinician, caregiver) and users are granted one or more roles rather than permissions directly. Widely used in healthcare, enterprise software, and increasingly in…

Shoulder Surfing

A form of visual eavesdropping where an attacker observes a person entering sensitive information such as passwords, PINs, or personal data by looking over their shoulder or from a nearby vantage point. Shoulder surfing is a particularly significant security concern for people…

Subtle Interaction(also: Discreet interaction)

A design approach in which interactions with a device or digital system are intentionally minimized in visibility, amplitude, or audibility to reduce social disruption, preserve user privacy, and support use in public or semi-public contexts. Subtle interactions include…

Surveillance

The systematic, focused, and often routine observation of people, their activities, or their data for purposes of influence, management, entitlement, or control. In accessibility and HCI research, surveillance is an analytical frame used to examine how monitoring technologies —…

Surveillance technology(also: Surveillance tech, Monitoring technology)

Technologies that collect, analyse, and track data about individuals' behaviours, locations, bodies, or communications. In disability contexts, surveillance technology raises justice concerns because assistive tools (computer vision for blind users, behaviour monitoring for…

Third-Party Tracking(also: 3rd-Party Tracking)

Tracking performed by a domain different from the one a user is visiting, typically via embedded scripts, ad networks, analytics tags, or browser extensions that phone home to external services. Third-party trackers aggregate browsing behaviour across many sites and can be…

Verifiable Credential(also: VC)

A cryptographically signed statement about a person, organisation, or object that a relying party can verify without contacting the issuer, following the W3C Verifiable Credentials Data Model. Verifiable credentials are the underlying data format used by identity wallets, and…

Visual Interpretation Service(also: VIS, Visual Assistance Service, Remote Sighted Assistance)

A service that provides visual information to blind and low vision users through human assistants, AI-powered tools, or a combination of both. Traditional visual interpretation services like Aira connect users via video call to trained human agents who describe visual…

Visual Privacy(also: Visual Information Privacy)

The safeguarding and management of sensitive visual information that could be shared or disclosed in everyday life, particularly through the use of assistive technologies and generative AI tools. For blind and low vision users, visual privacy encompasses multiple dimensions:…

Web Tracking(also: Online Tracking, Behavioural Tracking)

The collection of information about users across websites, typically by third parties, using techniques such as cookies, fingerprinting, tracking pixels, and network requests. Web tracking raises privacy concerns and can disproportionately affect disabled users who install more…

Zero-Knowledge Proof(also: ZKP, Zero-Knowledge Protocol)

A zero-knowledge proof (ZKP) is a cryptographic method by which one party (the prover) can convince another party (the verifier) that a statement is true without revealing any information beyond the truth of the statement itself. For example, a voter can prove they are eligible…

eIDAS(also: electronic IDentification, Authentication and trust Services, eIDAS 2.0)

A European Union regulation governing electronic identification and trust services across the single market. The revised eIDAS 2.0 regulation, adopted in 2024, requires each member state to provide a European Digital Identity (EUDI) Wallet to all citizens and residents by the…

ePrivacy Directive(also: Cookie Directive, EU Cookie Law, Directive 2002/58/EC)

A European Union directive that complements GDPR by specifically addressing privacy in electronic communications, including the use of cookies and similar tracking technologies. The directive requires websites to obtain informed consent before storing cookies on user devices.…