一颗爱心打败所有 AI，ChatGPT、豆包、Gemini全看不到

Core Points - The article discusses a visual illusion that has gone viral, which is being referred to as a modern Turing test for distinguishing between human and AI perception [1] - The illusion involves a picture where different viewers can see various shapes, such as a heart, a cow, or a coffee cup, depending on their perspective and cognitive interpretation [3][5][12] - The phenomenon highlights the differences in how humans and AI interpret visual information, with humans relying on experience and imagination while AI analyzes pixel data and geometric features [20][72] Group 1 - The visual illusion is created by the arrangement of black and white shapes, leading to misperceptions of movement and form [5][13] - Different viewers report seeing different shapes based on their cognitive biases and experiences, demonstrating subjective perception [16][18] - The article emphasizes that the brain actively constructs images from visual cues, rather than passively receiving them [12][20] Group 2 - AI models, such as ChatGPT and Gemini, struggle to identify the hidden shapes in the illusion, showcasing limitations in their visual processing capabilities compared to human perception [27][39][82] - The discussion includes examples of other visual illusions and how AI performs in recognizing them, indicating that while AI has improved, it still faces challenges with certain types of visual cues [86][96] - The article concludes by suggesting that the differences in visual interpretation between humans and AI may lead to future developments in AI's ability to understand human-like perception [100]