Text to image AI models bias toward clocks showing 10 10 time
It's Always 10:10: Reference Images Break a Bias That Prompts Only Dent
Computer Vision and Pattern RecognitionHuman-Computer Interaction
Summary
AI models trained to generate images often repeat common patterns seen in their training pictures, like clocks showing the time 10:10. This study shows that no matter what time you ask for, these models often default to 10:10, which is a common clock time in advertisements. The authors tested ways to reduce this bias by giving the AI clearer instructions and pictures, finding that adding a drawn reference clock helps the model show the correct times much more often. This improves accuracy and reduces the repeated 10:10 default.
What this means in practice
- •For ai developers: Improve text-to-image generation of time-specific objects by adding reference images to reduce common output biases.
- •For graphic designers: Use reference visuals with AI tools to create more accurate clock images at requested times for advertising or illustrations.
Authors
Luca Cazzaniga
Abstract
Text-to-image models appear to reproduce the habits of the photographs they learned from. Analog clocks are an extreme case: in advertising, watches almost always show 10:10, and generated clocks return to 10:10 even when another time is requested. We measure this bias and test three ways of overcoming it on 52 models available on the Magnific platform, with a replication on Higgsfield. Every image shows three identical clocks that must show 2:35, 6:50 and 11:20. The description of the object is fixed and only the request about the time changes: no time (A), the time in digits (B), the hand positions described by construction relative to the dial numerals (C), or the same description plus a drawn reference dial (D). Two AI readers read all 1,799 images blind from coded copies, with a third reader and the author settling disagreements (dial-level agreement 96.0% and 97.3%). With no time requested, 67% of the images have all three clocks at 10:10. On the 20 current models, all three clocks are correct in 34% of the images with digits, 30% with the hands described in words and 75% with the reference dial (D-B: +37 points, 95% CI +28 to +45); we found no evidence that describing the hands in words beats the digits (C-B: -4 points, CI -10 to +1). The replication on the 12 models shared by both platforms gives the same picture (B 54%, C 50%, D 81%). Writing the time reduces the bias but leaves two thirds of the images of the 20 current models with at least one wrong clock; adding a drawn reference raises full accuracy to three quarters and almost eliminates images entirely at 10:10. We release all images, prompts, raw readings and a script that recomputes every result.