Data Doctors: Why Google Photos search feels broken

Source Domain: wtop.com

Transition to AI-Driven Search: Google Photos now uses artificial intelligence for its search function, moving away from a simple keyword-based filing system to a more complex approach looking at faces, objects, scenes, text, and activities.
Confidence Score Threshold: The new AI searches assign a confidence score to photos and videos, and only those passing a specific confidence level appear in the results. This can cause users to miss items that were easily found before.
Challenges with Specific Searches: Specific searches like “dolphins in Greece” yield fewer results than simpler keywords like “dolphin” because they require the AI to confidently match multiple criteria simultaneously.
Video Search Difficulties: Videos face more issues due to AI’s need to interpret hundreds or thousands of frames, often excluding less obvious subjects or uncertain moments.
AI Integration: There’s no option to turn off the AI search. Instead, users are advised to adapt their search methods, employing broader searches and refining them with additional details.
Enhancing Search Effectiveness: To improve search outcomes, it’s recommended to start with broad, strong visual concepts and then narrow results by date or location, avoid combining multiple search ideas, add captions, create descriptive albums, and mark important items as favorites.

You may have missed