Recent reports indicate that Baidu has launched an AI social app called "Wen Xiaoyan," which is designed for real-time communication and interaction with AI virtual characters. This application is powered by the Wenxin large model technology and aims to establish emotional connections between users and AI.
User experiences suggest that while the app's features and functionalities are similar to other AI chat products, the use of realistic digital personas enhances the sense of interaction. Over the past year and a half, large models have evolved through three distinct phases, attracting more developers to contribute to AI ecosystem applications. This, in turn, is accelerating the practical application of AI technologies.
Pacific Securities highlights three key areas to watch in AI development this year:
1. The growth of AI applications targeted at B2B users and edge deployment.
2. Advances in multimodal technology, which can generate content across different formats and has significant commercial potential. Applications developed using multimodal models offer diverse interactive options, potentially expanding their user base.
3. Self-sustaining models, as increasing attention on domestic large models in China is prompted by developments like OpenAI's service interruptions.
In line with this trend, related publicly listed companies are making strides in the AI content creation space. For example, a subsidiary of Aotu Technology has introduced the Vormir generative tool platform, customized for use in Metabox digital live streaming environments. Additionally, InsightGPT, a multimodal large model developed by Insai Group, now offers functionalities such as text generation, image creation, video production, video editing, and audio synthesis. It has already begun to provide marketing materials like text, images, and videos for overseas service providers.