Since the end of last year, the generative AI wave represented by ChatGPT has swept the world, becoming the undisputed focus of the industry. While personal applications continue to innovate, industries are beginning to focus on how to apply generative AI to enterprise and industry scenarios. In this regard, Cao Hengkang, Global Vice President and President of Red Hat Greater China, stated that generative AI large language models are tools with extensive commonality and that exploring the integration of enterprise-specific scenarios to form enterprise-specific large language models is at the forefront of the industry. He mentioned that Red Hat’s recently launched OpenShift AI solution provides a platform for the implementation of generative AI in enterprises.

    Based on the mature functionalities and extensions of Red Hat OpenShift and OpenShift Data Science, OpenShift AI provides enterprises with a unified solution for training, monitoring, and managing AI models. Wang Huihui, Senior Director of Solution Architecture Department at Red Hat Greater China, stated that the development of AI applications first requires building a model, such as the ChatGPT large language model. Once the model is built, it needs to be trained with a massive amount of data. After the training is completed, the model needs to be deployed to establish corresponding AI applications, which can be deployed according to actual scenarios. Each of the above steps requires a significant amount of computing resources, and OpenShift has integrated resources from numerous AI ecosystem vendors, including NVIDIA, IBM, Intel, among others, which users can invoke these resources during the model training process.

    In addition, OpenShift AI already includes many open AI models. Developers can directly access industry-proven AI models on the platform and modify them according to their needs, avoiding starting from scratch. Wang Huihui explained that OpenShift AI includes the resources required for AI application development, such as AI applications, model training, data integration, and security control. All processes can be completed on one platform. After completion, the AI applications can be seamlessly deployed in various application scenarios, whether in the cloud, data centers, or at the edge, helping developers quickly develop and deploy AI applications.

    As one of the two important initiatives for Red Hat’s focus on the intelligent manufacturing field in the Chinese market in 2023, can this field become a breakthrough for the application of generative AI in industries? Regarding this question, Wang Huihui stated that Red Hat has already considered introducing generative AI into the intelligent manufacturing field and has established the platform. Currently, many intelligent manufacturing enterprises have already established big data analysis models, which serve as the preliminary preparation for the implementation of generative AI. The next step is to see the specific needs for generative AI application scenarios in the manufacturing industry. From the current perspective, the entire industry has already begun to explore and experiment.