Security, Privacy and Accessibility

Definition

Security, Privacy and Accessibility

Security protects systems, data and users from unauthorised access or misuse. Privacy governs how personal or sensitive data is collected, processed, retained and exposed. Accessibility ensures people with different abilities can use the system effectively. AI applications add specific concerns such as prompt injection, tool abuse, data leakage, unsafe outputs, consent, trace retention and accessible review of generated content.

Why It Matters

Security engineering includes threat modelling, least privilege, secure defaults, secrets management, dependency control, input validation and incident response. AI-specific threats include prompt injection, data exfiltration and unsafe tool use.

Privacy engineering covers data minimisation, purpose limitation, consent or lawful basis, retention, access control and privacy-preserving logs. Context and traces can contain more sensitive information than developers expect.

Accessibility and inclusive design ensure applications work for people with diverse abilities, languages, devices and contexts. AI output must not undermine otherwise accessible interaction.

Failure mode

The recurring risk is treating this concept as a label rather than an engineering decision with trade-offs, evidence needs and operational consequences.

Evaluation check

Ask what observable evidence would show that this concept has been applied well in a real system, repository or learning exercise.

Related Concepts

• Configuration, permissions and sandboxing
• Risk and responsible AI

Further study: OWASP, Top 10 for LLM Applications, and W3C, WCAG.