Trump's plan to redesign every .gov website leads to AI-designed horrors

2026-06-30 · Source: AI - Ars Technica · Field: Government & Public Sector — Digital Government & E-Government, Public Policy & Governance, Artificial Intelligence & Machine Learning · Depth: Fundamental Awareness, long

Summary

President Donald Trump's National Design Studio (NDS) launched the "America by Design" initiative to overhaul 27,000 .gov websites in three years using AI, following deep cuts to established government tech units like 18F. A year into its existence, NDS has shown limited progress, primarily launching single-page sites and facing accusations of claiming "false victories and overstated credit." Critics highlight its heavy reliance on AI, which led to errors like a six-toed child image on TrumpRX.gov, and significant failures in accessibility, including ADA non-compliance and shipping three megabytes of code for simple pages. Concerns also include the use of commercial visitor-tracking software on sites like ndstudio.gov and trumprx.gov, unclear data retention policies, and a lack of privacy impact assessments. Agencies are reportedly resisting NDS collaboration, and the requirement for NDS to update US Web Design System (USWDS) standards has been dropped.

Key takeaway

For government IT professionals considering AI for digital transformation, this case highlights critical risks. You must ensure robust human oversight for AI-generated content to prevent errors and maintain public trust. Prioritize comprehensive accessibility testing and privacy impact assessments before deployment, especially when handling sensitive user data. Relying solely on AI without established standards and inter-agency collaboration can lead to project failure and erode citizen confidence in digital services.

Key insights

The rapid, AI-driven overhaul of government websites by NDS has resulted in significant usability, accessibility, and privacy failures, hindering its core mission.

Principles

• Centralized, rapid tech overhauls often fail without institutional knowledge.
• AI-driven design requires rigorous human oversight and testing.
• Government digital services demand strict privacy and accessibility compliance.

Method

The NDS method involves using AI agents to generate complete website redesigns based on internal guidelines, aiming to accelerate the update of thousands of government sites.

In practice

• Avoid AI-generated images without human review to prevent errors.
• Prioritize ADA compliance and lightweight code for public-facing sites.
• Conduct privacy impact assessments before deploying new data collection.

Topics

• Government Digital Transformation
• AI in Web Design
• Web Accessibility
• Data Privacy
• US Web Design System
• Public Sector IT Governance

Code references

• uswds/uswds

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Policy Maker, IT Professional, AI Ethicist

Related on AIssential

• AI Agents Running the State — The "Agentic State" vision for AI-driven public services requires proactive risk mitigation through red-teaming its core assumptions.
• Pentagon boasts of using AI to write reports mandated by Congress — The Pentagon is rapidly adopting generative AI for critical tasks, risking accuracy and accountability.
• Fault Lines: Navigating Ethics and Responsible AI Where National Policy Meets Local Practice in Public Sector Transformation — Implementing responsible AI in UK public services faces critical national-local policy gaps and five interconnected challenges.
• The Web Is Being Made Accessible for AI, Not People — AI-driven infrastructure changes often prioritize technological systems over long-standing accessibility needs of disabled people.
• Amazon puts humans back in the loop as its retail website crashes from "inaccurate advice" that an AI agent took from an old wiki — AI integration without proper oversight and updated data can lead to significant system failures.
• White House Outlines National AI Safety and Innovation Strategy — The White House proposes a unified national AI strategy balancing innovation, user protection, and limited developer liability.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.