Researchers have investigated how color design affects the performance of artificial intelligence (AI) models in understanding maps. They created a benchmark with over 28,000 map-question pairs and tested 21 different AI models. The study found that disrupting the sequential order of colors on a map can significantly impair an AI's ability to compare or rank features. Additionally, reducing lightness contrast between colors also hinders performance. However, increasing contrast beyond what is necessary for human perception provides only minor benefits. These findings suggest that conventional cartographic principles are still relevant for designing maps that AI models can understand. --- Why it matters: This research matters because it highlights the importance of considering human-perceived visual cues when designing maps for AI models. Understanding how color design affects spatial reasoning in AI can inform the development of more effective and efficient AI systems, particularly those used in applications like geographic information systems (GIS). Source: https://arxiv.org/abs/2608.15736