ByteDance is secretly developing multiple products in the field of AI large models, including multi-modal digital human products, AI-generated image and video products. The company has made progress in the area of foundational large models, launching the first large language model "Doubao" and the multi-modal large model BuboGPT. In the AI application layer, ByteDance has established a new AI department Flow, and has introduced several AI dialogue products. As a video creation tool, Jianying is considered a promising product to integrate ByteDance's AI large models, and it will focus on text-to-video direction in the future.