According to PANews, Elon Musk said in a conversation with Stagwell Chairman Mark Penn that AI model training has exhausted real-world data and the way models are developed needs to be changed in the future.

Musk believes that synthetic data will supplement real data, and AI can achieve self-learning by generating and self-evaluating data. Technology giants such as Microsoft and Meta have already adopted this trend.

Gartner predicts that about 60% of data in AI projects will be synthetically generated in 2024. The advantages of synthetic data include cost savings, but there are also risks such as reduced creativity and output bias.