AI Transparency Lab

How Nala Assist uses intelligence without asking photographers to trust a black box. Structure first. Suggestions second. The photographer decides.

The trust problem

Bad AI creates homework.

If an assistant saves ten minutes and creates twenty minutes of verification, it failed. Nala is designed around a different principle: start where trust can be earned.

AI can be useful for photographers, but not when it hides the work, scores taste, or silently makes choices that the photographer must later audit.

Nala starts with simple, visible, easy-to-check help: groups, sequences, time gaps, similar frames, duplicate candidates and technical signals. These are structures and cues, not verdicts.

Nala principles

Signals, not hidden judgment.

Nala’s role is to prepare the Light Table before the photographer begins making decisions.

September preview direction

Nala starts where trust can be earned.

The private September preview focuses on useful organization: grouping by EXIF, time, camera and lens; burst candidates; focus stack candidates; similar-frame groups; duplicate and near-duplicate detection; multi-folder and multi-SD card consolidation; and multi-camera timeline reconstruction.

Multi-photographer project structuring is a later direction. It matters, especially for events, weddings, media and productions, but it belongs to a deeper version of Nala once the simpler structures are trusted.

What Nala will not do

No silent deletion. No taste score. No magic best-photo claim.

Nala does not replace the photographer’s eye. It does not pretend that taste is a number. It does not hide images as a shortcut.

The long-term ambition is not to make the photographer supervise an algorithm. It is to make the shoot easier to understand before the photographer begins.