• Digitized archives still exclude blind and low-vision users
• Many images lack meaningful alt-text or rely on minimal, generic captions
• Alt-text is not just a technical requirement but part of historical justice
• Rich descriptions also benefit other groups
  (e.g., neurodivergent readers)

• Alt-text translates visual evidence into language
• Production is labor-intensive; often postponed or dropped entirely
• Requires:
  • Domain knowledge (history, archives, local context)
  • Accessibility expertise (WCAG, screenreader use)
  • Editorial judgement (what is salient, what is harmful)

• Multimodal models can caption images at scale
  (e.g., GPT-4o mini, Gemini, Llama, Qwen)

• Promises

  • High coverage, fast throughput, low marginal cost
  • Potential to fill the gap for large collections

• Risks

  • Hallucinations, misrecognitions, and subtle biases
  • Uncritical reproduction of harmful terminology from sources or training data

• RQ1 — Feasibility
  • What coverage, throughput, and unit cost can current VLMs achieve for WCAG-aligned alt-text on a heterogeneous heritage corpus?
• RQ2 — Relative quality
  • How do humanities experts rank model outputs?
  • Which error patterns recur in ‹best› alt-texts?

• Media Types: photographs, maps, drawings, objects, diagrams, ephemera
• Time Periods: from prehistoric to contemporary Basel
• Complexity: multi-part figures, text-heavy items

• Four vision–language models via OpenRouter:
  • Google Gemini 2.5 Flash Lite
  • Meta Llama 4 Maverick
  • OpenAI GPT-4o mini
  • Qwen 3 VL 8B Instruct
• Selection criteria:
  • Mix of proprietary and open-weight models
  • Similar cost range and context window
  • Multimodal and multilingual (incl. German)

• Fixed, WCAG-aligned system prompt
  • Short, neutral, factual descriptions
  • No “image of …”, no alt= wrappers, no emojis
  • Length targets
    • 90–180 characters
    • up to 400 for complex charts
  • Rules for portraits, objects, documents, maps, event photos

• User prompt
  • Concise, structured metadata (title, description, date, era, creator, publisher, source)
  • Image URL at the end

• Participants: 21 humanities scholars

• Instructions:

  • For each of the 20 images, rank 4 model outputs
    from 1 (best) to 4 (worst)
  • Focus on WCAG-aligned criteria:
    • Core visual content, no redundancy
    • Salient features and visible text
    • Context only when it aids understanding
  • Factual accuracy and bias not foregrounded but may influence judgement

• Coverage

  • All four models produced non-empty alt-text for all 20 images → 100% coverage

• Latency and throughput

  • ≈2–4 seconds per item; ≈0.24–0.43 items/s
  • Qwen 3 VL 8B: fastest; GPT-4o mini: slower but stable

• Cost

  • ≈1.8×10⁻⁴ to 3.6×10⁻³ USD per description
  • Sub-cent cost even for multiple candidates per image

• technically and economically feasible at collection scale

• Rank distributions

  • GPT-4o mini and Qwen receive more first-place and fewer last-place ranks
  • Gemini and Llama perform slightly worse descriptively

• Statistical tests

  • Friedman test: χ²(3, N=20) = 6.02, p = 0.11
  • Kendall’s W ≈ 0.01 → very low agreement across tasks
  • Pairwise Wilcoxon tests (Holm-corrected): no significant differences

• no clear ‹winner›; relative model quality varies by image

• Operationally viable (RQ1)
  • High coverage, low latency, negligible cost
  • Easy to integrate into existing pipelines
• Epistemically fragile (RQ2)
  • Errors are uneven and task-dependent
  • Biases and omissions require domain-sensitive review

• Single institutional corpus; 100-item dataset, 20-item survey subset
• Four models at one point in time (Oct 2025)
• Future directions:
  • Larger and more diverse participant pools, including blind and low-vision users
  • Comparative studies of different prompt and review strategies
  • Integration with authoring tools and cataloguing workflows

• Recommended use:
  • VLMs as drafting aids for trained editors
  • Human-in-the-loop workflows with:
    • Editorial guidelines for sensitive content
    • Checks for bias, factual errors, and omissions

• For DH Teams and GLAM Institutions:

  • Start with limited, well-documented pilot runs
  • Prioritise sensitive or high-impact items for manual review
  • Develop internal style guides for alt-text and metadata
  • Log model version, prompts, and provenance for transparency

• For Model and Tool Builders:

  • Support context-aware prompting (metadata + image)
  • Provide better controls for bias and harmful content

• Do not use VLMs as fully automated captioners
• Treat alt-text as:
  • An interpretive, historiographical act
  • An ethical responsibility toward diverse audiences
• Effective adoption requires:
  • Clear editorial policies and sensitivity guidelines
  • Human-in-the-loop review with domain and accessibility expertise
  • Transparent, reproducible workflows and open data

→ Accessibility with AI is a design and governance choice