MAJORFailure Reference

Google Veo Hallucinated Text — Pre-Generation Risk Reference

Technical Classification

Glyph Synthesis & Semantic Adherence Failure

Like all video diffusion models, Google Veo struggles with in-scene text rendering. Signs, labels, screen UI, and t-shirt graphics frequently produce shapes that resemble letters but don't form readable words. Veo 3 marginally improved on Gen-3 here but still fails on text >5 characters and on cursive / stylised typography. This is a Glyph Synthesis Failure, and because Veo is a Google product with the most rigorous support process of the major platforms, well-documented requests approve at a higher rate than competitors.

How to identify this failure

  • Storefront sign reads as garbled non-letters
  • Same word changes characters frame-to-frame
  • Subtitles show wrong words or mirrored glyphs
  • T-shirt or product label text is unreadable
  • Numbers render incorrectly even when explicitly prompted

Real generation examples

Prompt used

"Coffee shop storefront with a neon sign reading OPEN 24 HOURS"

Failure observed @ 0:00 - 0:08

Sign reads "OPΞИ 24 HOIIRS" with mirrored second-letter forms

Prompt used

"Whiteboard with the words BRAINSTORM SESSION written on it"

Failure observed @ 0:02

Whiteboard text is illegible glyphs; letter shapes morph between frames

Documentation strength

If you need to escalate

HIGH — Google's support process is the most structured of major AI video platforms; well-documented text-failure cases approve consistently.

AVA is a pre-purchase prevention tool, not a post-purchase recovery tool. Platforms generally do not guarantee credit refunds for output-quality failures; goodwill credits are at each platform's discretion. The strength rating reflects how well-formed your support ticket can be, not a promised outcome.

Prevention + documentation steps

  1. 01

    Score your prompt before you generate

    Run your prompt through AVA's pre-flight scoring against the Glyph Synthesis & Semantic Adherence Failure pattern. Green light = generate. Yellow/red = rewrite using the suggested fix before you commit credits.

  2. 02

    Capture Generation ID + timestamp if it failed anyway

    Find the Generation ID in the URL or share link. Note the exact time when the Glyph Synthesis & Semantic Adherence Failure first appears (e.g. "failure first visible at 1.2s"). Timestamped evidence is significantly stronger than a general complaint.

  3. 03

    Use the correct technical term in your support ticket

    Describe this failure as "Glyph Synthesis & Semantic Adherence Failure". This term maps to a recognised internal workflow in the support system and routes the ticket to the right team.

  4. 04

    Submit via the correct support channel

    Runway has no direct email intake. Pro+ plan: open the in-app AI Assistant (help widget bottom-right of app.runwayml.com), describe the failure with the technical term, attach evidence. Free/Standard plan: human support isn't available — your channel is Discord #community-help with @On Call - Moderators.

Frequently asked questions

Will Google Veo refund credits for garbled text?

Yes, when your prompt specified exact text and the output produced glyph noise. Cite "Glyph Synthesis Failure" plus the Generation ID, quote both the prompt text and what was actually rendered, and attach the PDF audit report.

Why does Veo hallucinate letters?

Video diffusion training emphasises natural imagery; text is sparse in the training set and rarely temporally stable, so the model produces letter-shaped artefacts rather than coherent typography.

How do I render text correctly on Veo?

Use short text (1-3 letters), avoid cursive, use large display fonts. Veo handles text marginally better than Runway Gen-4 but still fails > 5 characters most of the time. AVA flags any prompt with quoted text as high-text-risk before generation.

Score your prompt

Score your prompt against this failure mode in 30 seconds

Paste your prompt and the platform you intend to use. AVA returns a red/yellow/green score against this specific failure mode plus a concrete rewrite if the risk is high.

AVA Pro · founders' round

$50 for 6 months of unlimited scoring across all failure modes + personal failure-history dashboard. Locks in $13/mo grandfathered after.

Claim $50 founders

Related failures across models

If you’re seeing this failure, you may also encounter these on other models:

Pick a different tool for Veo failures

Some prompt shapes will keep failing on Veo. Routing those shots to a different vendor is the cheapest fix.