AI Image Policy
Why this site uses generated images
Lokavidu covers subjects that are difficult to photograph: private amulets not in public collections, monks who lived before photography, consecration ceremonies closed to outsiders, sacred objects that may not be handled freely, and temples whose interior appearance is undocumented in English-language archives. Where no licensed photograph exists and the subject is visible, a correctly classified generated image teaches more than a blank space.
AI generation is the default when no usable source image exists. It is not a substitute for a strong primary image that is legally available.
Image classes
Every AI-generated image on this site is assigned one of four classes. The class appears in the visible caption that travels with every image.
- Representative visualization
- Shows the documented characteristics of an amulet type or object class. Not a photograph of any individual object.
- Historical reconstruction
- Interprets a past place, person, ceremony, or process using cited sources. Not a primary record.
- Editorial illustration
- Conveys a concept, atmosphere, or relationship. Makes no claim of documentary accuracy.
- Explanatory diagram
- Clarifies parts, sequence, geography, comparison, or ritual structure.
What AI images are never allowed to do
Generated images on this site may not be presented as:
- a photograph of an authentic historical amulet;
- evidence of provenance, age, material, mold, temple origin, or authenticity;
- an actual historical photograph of a monk or ceremony;
- a faithful record of a temple's present appearance unless verified against current references;
- a primary source; or
- evidence supporting a factual claim.
Generated visuals may explain documented information. They may not become the documentation.
Living monks and verified likenesses
Photorealistic likenesses of living monks are not generated from text alone. Where a real licensed photograph is available, it is used. Where no verified likeness exists, the image uses an obviously interpretive medium and does not fabricate documentary accuracy.
Disclosure
Every AI image carries a visible caption that names its class. Disclosure is not buried in metadata, alt text, or a global policy page. The caption travels with the image wherever it appears. An image without its disclosure caption is not published.
Production model
The default production model for new images is gpt-image-2 (high quality). All generated files are processed through a Sharp pipeline: auto-rotate, maximum 1600px width, WebP at 80% quality. EXIF and AI metadata are stripped before placement.
Source priority
When a source image exists, it is used ahead of generation: rights-cleared photographs, public-domain or licensed archival images, and original photographs commissioned for Lokavidu take precedence over any generated image in all contexts.
Questions
Questions about specific images or this policy can be sent to the address on the About page.