A new study published in Nature reveals that AI models can now generate highly realistic synthetic images that are indistinguishable from real photographs to the human eye. The research team developed a novel architecture that significantly improves the coherence and fine detail of generated images, addressing common flaws like distorted hands or inconsistent lighting. While the technology promises advancements in creative fields and data augmentation for research, it also raises significant concerns about the proliferation of deepfakes and misinformation. The authors call for the development of robust detection tools and public education to mitigate these risks. Read the full article at https://technologyreview.com/2024/05/15/ai-image-synthesis-breakthrough.