Designers are already making security decisions. They are just not doing it deliberately.

Every form field, every permission toggle, every sharing feature carries a data decision. Sensitive information travels, and the question is how and to where. When AI tools generate code from design specifications in seconds, the security implications of those choices scale instantly.

Stanford research from 2025 found that teams using AI assistants produced code with security vulnerabilities at a rate of 40%, compared to 27% without. That gap exists because AI implements what is specified. The vulnerability is upstream, in the design.

Angelos draws from direct experience designing a cybersecurity platform where security and usability must be strictly interlocked, and from documented breaches where a single design choice determined the impact. The talk concludes with a practical framework for integrating security thinking into everyday design workflows.