Alt Text That Works for Humans and AI: A Writer’s Guide

Accessibility standards and generative AI tools shape best practices for alternate text on images.

Alt text provides a concise description of the image. Screen readers, which visually impaired readers rely on, speak the text. Generative AI tools also use alt text to interpret the image.

Example: “Network diagram showing three tier architecture with web, application, and database layers.”

Guidelines:

  • Screenshots in procedural steps:
    Provide descriptive alt text only. Human readers and AI agents use context and position to interpret the screenshot. Do not include visible descriptions above screenshots. This impedes readability and interrupts procedural flow.
  • Complex diagrams:
    Include a brief description above the diagram to orient all users before they interpret the visual. Also provide alt text.

How do you use alt text in a document?

AI Won’t Replace Technical Writers. But Technical Writers Who Use AI Will Replace Those Who Don’t.

Technical writers face a paradox in 2026: AI promises to replace us while simultaneously becoming one of our most helpful tools. Here’s how I’ve navigated that paradox to reduce project completion time by up to 40% without sacrificing quality.

My AI workflow:

I use ChatGPT and Microsoft Copilot as initial generators, not finished product creators. I provide the topic and objective and let AI produce an initial draft. Then the real work begins.

What I bring that AI can’t:

  • User advocacy: I anticipate the questions users will ask. The stumbling blocks. The missing context. AI generates information. I generate understanding.
  • Context reading: Copilot regularly suggests edits that can be contextually wrong. Recognizing that difference requires human judgment.
  • Style guide enforcement: Consistency across documentation requires knowing the guide AND knowing when to apply it. AI doesn’t do this reliably.
  • Fact checking: AI sounds confident when it’s wrong. I verify everything.

The truth about AI in documentation:

It’s a powerful first draft generator and a useful reviewer. It is not a technical writer.

What’s your experience using AI in your documentation workflow? I’d love to hear how other writers are navigating this issue.