In 2022, with the popularity of artificial intelligence technologies such as autonomous driving, ChatGPT, and AI painting, AI will begin to enter the public context. In Baidu, Tencent, Alibaba Dharma Research Institute and MIT Technology Review’s predictions on the top ten technologies in 2023, AI technology has also been given an important chapter.

1. AIGC

In 2022, AIGC (AI generated content, artificial intelligence self-generated content) will be popular all over the world. From Google, Meta, Microsoft, to Stability AI, Jasper, and OpenAI, they will all lay out the AIGC track, and the limelight of the industry will never change. According to the analysis of “Generative AI: A Creative New World”, from text, images, to video, audio, AIGC has great advantages in terms of efficiency and creativity. AIGC has the potential to generate trillions of dollars in economic value. By 2025 Annual generative AI will account for 10% of all generated data.

Thanks to the breakthroughs in basic research of large models in recent years, especially deep learning, generative AI can generate brand new digital video, image, text, audio or code from massive training data. For example, ChatGPT, which can be used for question and answer, text writing, and code writing, and AI painting software for various Vincent diagrams.

With the support of multiple factors such as open source models and falling computing power costs, the commercialization of generative AI has begun to take shape. At present, the works generated by AI painting have begun to be used in many fields such as advertising marketing, video production, and game model production. In the future, with the further improvement of generative AI in terms of cognitive intelligence and generation controllability, generative AI will also enter a period of application explosion, which will greatly promote the production and creation of digital content.

2. Autonomous Driving

Autonomous driving is undoubtedly the most complex AI application scenario today. More than 250 automakers, truck manufacturers, start-ups, and transportation and mobility-as-a-service providers around the world are working on this challenge to develop autonomous vehicles.

In 2023, autonomous driving will gradually enter the urban scene, and there are still many technical difficulties to face: whether it is the difficulty of sensing complex environments or processing massive data, it will greatly increase. At this time, traditional small models cannot meet the requirements of high-level autonomous driving. The industry has begun to introduce large-scale model technology to allow self-driving cars to effectively expand semantic recognition data, greatly improve the efficiency of solving long-tail problems, further enhance the generalization ability of self-driving perception, and adapt to more travel scenarios.

3. Digital Human

Market analysis agencies predict that by 2026, the scale of China’s AI digital human market will reach 10.24 billion yuan, and the digital human market will show a trend of rapid growth. The “Digital Human Industry Development Trend Report (2023)” released by Tencent pointed out that the current digital human production tools and solutions are relatively complete, and digital human has been widely used in the fields of enterprise services, film and television animation, and cultural media.

Thanks to the addition of AI technology, the development cycle and cost of digital humans have been greatly shortened and reduced; at the level of intelligence, digital humans mainly rely on NLP for text-driven, and based on the development of large NLP models, the degree of intelligence of digital humans is increasing day by day. At the application level, at present, digital humans are widely used in digital marketing, entertainment and other fields. With the advent of the era of true Internet, digital humans will become an important element and new entry point.

4. Robots

As the world ages, the importance of robots becomes more apparent. With the intervention of technologies such as AI, big data, and cloud computing, the performance of robots in perception, decision-making, and execution has been greatly improved.

Among them, in terms of perception, tactile perception is the focus of research in the field of robot perception completion. Benefiting from breakthroughs in flexible materials, tactile sensing technology has been developed and tested in robotic hands, tactile gloves, health testing equipment, and smart cockpits. For example, the flexible gripper installed on the end effector of the collaborative robot can sort and grasp objects with complex shapes and materials, just like human hands.

In 2023, with the further decline of robot ROI, various robots such as mobile robots and collaborative robots will begin to enter factories and homes, taking on the important task of reducing costs and increasing efficiency.

5. Scientific Computing

AI technology has become an important auxiliary force for scientific research, upgrading the multidisciplinary research paradigm. In 2022, AI for Science will receive more and more attention-the success of models such as AlphaFold has made people see that AI technology can have a huge impact on scientific computing and is changing the research paradigm of many disciplines. Baidu Research Institute believes that by introducing AI technology, researchers have developed scientific computing tools to solve problems that are too complex and difficult to solve in traditional scientific computing, and improve system modeling and analysis capabilities.

Therefore, they believe that more powerful scientific computing tools will emerge in the future, promoting AI technology to become an important auxiliary force for scientific research, and embodying its own value in basic sciences such as physics, chemistry, biology, materials science and drug research and development.