Generative AI models face risks of 'model collapse' due to reliance on synthetic data, leading to potential degradation. Researchers emphasize the need for high-quality, human-generated data and renewed focus on robotics and embodied AI.
🔗 in comment 
Opinion Artificial intelligence
AI can learn a lot from its biological predecessors
Synthetic data risks ‘model collapse’ but may spur renewed interest in robotics
John Thornhill