27 September 2026
25 new posts across 3 design and research sources — 6 worth your time.
When Should You Disclose AI Use? The PACED Framework
Explores how user reactions to AI disclosures vary based on audience, context, and the degree of AI involvement. It introduces the PACED framework to help teams navigate transparency requirements, such as those mandated by the EU AI Act starting in August 2026.
Nielsen Norman · 27 September 2026 · Read the original →
The variance problem in AI products
Discusses how inconsistent performance speeds in AI products damage user trust more than consistently slow response times. Written from firsthand engineering experience, it highlights the need to manage user expectations when system latency fluctuates.
UX Collective · 27 September 2026 · Read the original →
AI Can Build Your UI. Now Figma Can Tell It When It Screwed Up.
Details a new Applitools integration that uses Figma designs as visual baselines to automatically detect discrepancies introduced by AI coding agents. This helps bridge the gap between design intent and AI-generated production code.
Web Designer Depot · 27 September 2026 · Read the original →
Nobody starts as an accessibility expert
Addresses the intimidation designers face regarding accessibility compliance and argues for shifting accessibility considerations to the far left of the project timeline. It encourages designers to take ownership of compliance early rather than treating it as a post-launch development task.
UX Collective · 27 September 2026 · Read the original →
Apple’s Foldable iPhone Has a Left-Handed Problem
Analyzes the UX implications of Apple moving key physical controls to the side of its new foldable iPhone, which creates usability challenges for left-handed users. It prompts interface designers to reconsider physical ergonomics and handedness in hardware-software integration.
Web Designer Depot · 27 September 2026 · Read the original →
How AI Works and How Users Think About It: Study Guide
A curated study guide compiling research on how artificial intelligence models function and how users conceptualize them. It emphasizes that users do not need to fully understand AI mechanics to use them effectively, but designers must design for their mental models.
Nielsen Norman · 27 September 2026 · Read the original →
More AI means more frameworks and more debugging.
6 stories, every Sunday
Published here every week. Follow by RSS to get it as it lands.