Dr. Xiao Song, Executive Vice President of Siemens Global and Chairman, President and CEO of Siemens China, represented Siemens at the 2024 World Artificial Intelligence Conference (WAIC) Industrial Development Plenary Session. At this world-class AI technology event, he shared his unique thoughts on "reshaping the future of industries with AI and turning imagination into productivity".
The full text of Dr. Xiao Song's speech：
Good afternoon everyone! I am delighted to represent Siemens at the World Artificial Intelligence Conference.
I believe everyone has experienced the beauty of generative AI. With just a few words, one can create a realistic and vivid image through the use of text images and videos.
Imagine one day, AI will generate not just a video, but an immersive, highly simulated, and physically compliant virtual space. Being immersed in it, people can easily handle complex tools and tasks with the assistance of AI, and freely engage in invention and creation. Devices and production lines in real space can collaborate more efficiently and autonomously with the assistance of AI, turning virtual world designs into "physical objects". That is to say, using natural language, ordinary people can create any product they can think of!
What kind of wonderful experience will this be? The digital and real world are becoming more deeply integrated, bridging the gap between humans and machines. Products that took years to develop in the past may appear in the future in just a few hours or even seconds. Human creativity is greatly unleashed, and production efficiency is greatly improved.
This is the "industrial metaverse" envisioned by Siemens. The explosive growth of AI will accelerate the realization of the "industrial metaverse", enrich its connotation, and ultimately reshape the entire industrial value chain!
As early as over 50 years ago, Siemens embarked on the journey of industrial AI, accumulating profound experience in machine learning, visual intelligence, anomaly detection, knowledge graphs, neural networks, and other fields. It has over 1500 AI experts and more than 3700 AI patent applications worldwide, ranking first in Europe.
We are well aware that industrial scenarios have multiple interactive links, significant industry differences, high complexity, and low fault tolerance. If AI wants to move from consumption to industry, it must deeply integrate with industrial scenarios, connect the digital and mechanistic channels, and achieve a leap in productivity with safe, reliable, and trustworthy industrial grade AI.
Siemens focuses on more than 40 key industries, serving over 400000 customers and accumulating rich on-site experience and industry know-how. One third of the factories worldwide are using our controllers. Siemens is also one of the top ten software companies in the world. In the automotive industry, 9 out of the top 10 vehicle manufacturers in the world use Siemens industrial software. These are the foundations of industrial AI innovation.
Not long ago, we showcased the world's first generative AI for industrial environments at the Hanover Industrial Fair, Siemens Industrial Copilot, which is based on Microsoft's big language model and Siemens' industry experience. Engineers only need to input natural language to generate complex automation code, greatly reducing programming barriers and shortening development time. This tool has been put into use on Schaeffler's production line.
It is estimated that the global manufacturing industry generates approximately 1800+PB of data per day. How to identify and decode this vast amount of data is another major constraint for the implementation of AI in the industry. Siemens OT/IT fusion technology can efficiently collect and integrate multimodal data, and understand and apply it through models.
We are developing and training an "Industrial Time Series Data Basic Model" based on 33 billion high-quality time series data from manufacturing, buildings, transportation, and other fields, to achieve various applications such as trend prediction and anomaly detection. In the future, we will work together with more partners to accelerate the generalization and scaling of industrial AI through a new paradigm of "basic models+intelligent applications".
While everyone is discussing that '2024 is the year of AI scenarios', a large number of AI and scenario combinations have already appeared at Siemens' Chengdu Digital Factory. So far, we have implemented nearly 100 AI projects, with the vast majority coming from innovation by frontline workers and technicians, achieving a continuous 5-year reduction in factory manufacturing costs and a process quality of up to 99.999%. This is the mutual achievement of human intelligence and AI!
As Roland Busch, CEO of Siemens, said on this stage three years ago, no country or organization can achieve digital transformation alone.
Entering the AI era, only through symbiosis can we achieve mutual benefit!
Siemens is already deepening cooperation with domestic and foreign companies such as Nvidia to accelerate the layout of the industrial metaverse. Two years ago, we launched the open digital business platform "Siemens Xcelerator", which will become a "shared platform" for us to explore industrial scenarios, connect industrial data, and accelerate the process of industrial AI generalization and scaling with ecological partners from various industries.
Technological innovation has propelled human industrial civilization from steam engines to electrification, automation, digitization, and intelligence. Now, we are at the forefront of a new round of industrial revolution.
Let's work together to unleash the infinite possibilities of the future with the great innovation of AI, realize the 'industrial metaverse', and turn imagination into productivity!