Designing with Standards for the AI Era
The importance of interface standards in AI and lessons from history.
1A Historical Perspective
In the 1990s, web browsers operated by their own rules. Netscape, Internet Explorer, and Opera rendered the same page differently, each using proprietary tags. This chaos forced designers to build the same site multiple times.
2The Birth of Standards
Jeffrey Zeldman and colleagues recognized this chaos as a choice. They launched the Web Standards Project, convincing browser makers to support common technologies. This movement made the web a universal platform.
3AI's Wild West
Today, AI interfaces face similar chaos. Platforms like ChatGPT, Claude, and Gemini approach content processing, sourcing, and user interaction differently. Like the browser wars, lack of standards creates problems.
4Zeldman's Legacy
Zeldman's real achievement wasn't writing specifications but convincing an industry to believe in shared standards. His 2003 book 'Designing with Web Standards' explained why separating structure, presentation, and behavior matters.
5Costly Repetition
Standard-less AI interfaces incur costs like the 90s web did. Inconsistent experiences, user distrust, and quick obsolescence emerge. Each team must develop its own solutions without shared conventions.
6Design Layers
Zeldman's core principle: separate structure, presentation, and behavior. This applies to AI interfaces too. Model logic, user interaction, and interface design should be handled in distinct layers.
7Industry Organization
Web standards emerged from grassroots efforts by designers and developers, not top-down mandates. AI needs similar collaboration. Independent practitioners must develop a common language.
8Education Gap
Many designers today don't know how standards were won. This historical amnesia increases risk of repeating mistakes. Design education should emphasize the evolution of standards.
9User-Centric Approach
Users always benefited most from standards. Consistent AI interfaces will build trust. Common rules are needed for sourcing, permissions, and transparency.
10Time Window
As Dan Maccarone noted, we have a brief window to shape standards. As AI ecosystems mature, change becomes harder. Acting now can prevent future chaos.
11Practical Steps
To create AI interface standards: 1) Industry leaders must convene 2) Document best practices 3) Share open-source solutions 4) Systematize user feedback.
12Enduring Legacy
Zeldman's struggle showed us: Technologies are temporary, principles endure. In the AI revolution, human-centered, accessible, sustainable standards will be our most valuable legacy.
Original article: UX Collective - Medium — AI Döneminde Web Standartlarıyla Tasarım