Group 1 - A research team from the University of California, Los Angeles has developed a new optical-based image generation model that significantly reduces energy consumption for generative AI, paving the way for a more sustainable future [1] - Current generative AI systems face scrutiny due to their high energy consumption and environmental impact, requiring substantial electricity and water resources for cooling data centers, raising concerns about long-term sustainability [1] - The new model utilizes an innovative computing paradigm that combines digital encoding with optical decoding, allowing for a single-step image generation process that enhances efficiency and reduces computational burden [1] Group 2 - The optical system has demonstrated the capability to perform complex generative AI tasks at scale, potentially leading to faster and more energy-efficient AI systems that can be integrated into everyday devices [2] - Testing on multiple datasets has shown that the optical model can generate images such as handwritten digits, clothing, butterflies, and faces with significantly lower energy consumption compared to traditional methods [2] - The system offers unique security and privacy advantages, and its compact size and low power consumption enable integration into smart glasses, augmented reality, virtual reality headsets, and mobile devices, promoting localized real-time content generation and supporting the development of edge computing [2]