WAIC officially concluded, but the accelerated implementation of AI still faces new challenges

After three consecutive days, the 2021 World Artificial Intelligence Conference (WAIC) officially closed.

Shedding the mysterious cloak of artificial intelligence technology and its industrial development, and after experiencing the pain of the bubble bursting and the setbacks faced by many unicorns in listing, this artificial intelligence conference has become more pragmatic.

The strength of chip and car manufacturing hard technology is highlighted

If we have to point out the biggest difference of this conference, it would be the highlighting of hard technology strength, which can be said to be the well-deserved first choice.

First of all, chip manufacturing has become a major symbol of the hard technology strength of this conference. According to incomplete statistics from Sina Technology, there are more than 20 chip companies participating in the exhibition, occupying nearly one-sixth of the entire exhibition area. From the perspective of chip companies participating in the exhibition, except for Qualcomm, an overseas company, they are mainly traditional domestic listed companies such as Fudan Microelectronics and Fudan Microelectronics, as well as new local chip manufacturing companies such as Cambrian, Tianshu Zhixin, and Suiyuan Technology.

In addition to chip manufacturing, at this summit, leading technology companies such as Alibaba, Ant Financial, Meituan, JD.com, and iFlytek have demonstrated their hard power and launched consumer electronic products such as chip modules, privacy computing hardware products, network servers, unmanned delivery vehicles, drones, and smart watches. A development path for Internet companies from soft to hard is constantly expanding. "Why are these products becoming more and more like Huawei?" asked an audience member at the exhibition area.

In fact, since hardware products are more tangible and more easily standardized than software products, the "model of providing services through software but charging fees by selling hardware" has been increasingly recognized by the outside world in recent years. Unlike Huawei's "from hard to soft" route, the trend of mainstream Internet companies from soft to hard was further highlighted at this exhibition.

In addition, at this summit, Alibaba, 360, Pony.ai, AutoX and even SenseTime all exhibited their smart cars. Smart cars, which are ultra-large hardware intelligent electronic products, not only cater to the current craze for car manufacturing, but also highlight the determination of technology companies to transform from soft to hard.

Digital RMB sweeps inside and outside the exhibition hall

The emergence of the digital RMB is undoubtedly the biggest feature that distinguishes this World Artificial Intelligence Conference from previous ones.

In April last year, the central bank announced the start of the first batch of digital RMB pilots. Shenzhen, Suzhou, Beijing, and Chengdu became the first batch of digital RMB pilot cities to take the lead in implementing the exploration. Subsequently, the promotion of digital RMB in various parts of the country has been one wave after another. According to incomplete statistics from Sina Technology, in the first half of this year alone, 11 digital RMB red envelope promotion activities have been carried out across the country, with a total of more than 100 million yuan in red envelopes distributed.

At this conference, three of the six digital RMB operating institutions designated by the central bank, namely Bank of Communications, Industrial and Commercial Bank of China, and China Construction Bank, all participated in the exhibition. Judging from the exhibitions of each bank, digital RMB has become their main promotion object. Not only digital RMB ATMs were exhibited on site, but also digital RMB media products such as smart watch wallets, coin wallets, and electronic cards were exhibited.

Surprisingly, near the elevator building of a large supermarket near the exhibition hall, the QR code of the digital RMB has been placed in front of the store’s payment counter, and users can choose to pay and settle through digital RMB according to their personal wishes.

This year's artificial intelligence is more down-to-earth

In addition to the frequent appearance of smart cars, chips and digital RMB, scenes such as virtual digital humans, autonomous driving test rides, AR+AI immersive experiences, AI portraits, robot performances, etc. also appeared at this conference with varying frequencies, attracting most of the audience to stop and appreciate the experience. However, as frequent exhibitors at previous years' conferences, they are not a feature of this year's exhibition.

Overall, the application experience of this year's conference venue and surrounding intelligent scenes has been further enriched, and various companies are also competing to showcase the results of their digital transformation.

An obvious trend is that AI applications are already within reach. During this year's conference, many cutting-edge technology products and actual application cases were not only presented at the technical level, but we also saw more practical application scenarios. At the same time, the widely mentioned digitalization and intelligent transformation of industries is motivating the AI ​​industry to quickly move to the next stage.

As Xia Huaxia, vice president and chief scientist of Meituan, said at the Artificial Intelligence Industry Development Summit Forum held on the afternoon of the 8th, after more than 60 years of development, and with the industry's continuous promotion of new AI infrastructure since this year, this has indirectly reflected that AI has reached the stage of implementation and becoming a general infrastructure.

During the AI ​​conference, Xu Yingtong, president of Huawei's Ascend computing business, announced that the company will complete the construction of more than 20 AI computing centers by the end of the year. At the same time, Tencent's senior executive vice president and CEO of the Cloud and Smart Industry Group Tang Daosheng also said that Tencent's Yangtze River Delta AI Supercomputing Center, which was built with an investment of 45 billion yuan, will also be put into use in the near future.

As artificial intelligence computing centers invested heavily by companies such as Huawei and Tencent are put into use one after another, the improvement of the infrastructure layer is accelerating the further implementation and development of upper-level AI applications.

Bridging the digital divide: AI technology faces trustworthy challenges

With the booming development of digital intelligence, new imbalances are also emerging. Some elderly people are blocked from using smart devices because they do not know how to use input methods; people in remote areas who are not yet connected to the Internet are losing the ability to understand and master digital technology along with developed areas; traffic-based platform companies have further expanded their competitive advantages by using the value mined from massive data, while small and medium-sized enterprises and individual consumers are blocked from the door.

How to bridge the growing digital divide has become a topic worth discussing. Improving people’s trust and acceptance of AI has become an important issue affecting the development of artificial intelligence and its industry.

At this AI conference, all participating Internet giants exhibited and introduced their company's goal plans in digitalization for the elderly and low-carbon development, further demonstrating their corporate social responsibility. At the same time, during the theme forum speeches promoted by the parallel exhibitions, discussions on AI ethics and governance, social responsibility and safety and reliability were also put on the agenda many times.

However, the next stage of development of AI technology will inevitably face challenges. On the technical level, how to make AI more robust and more replicable is a challenge facing trusted AI. At the same time, the explainability of AI is also becoming a key factor restricting the development of AI technology and its industry.

"When we help companies with flexible production and design, we will design products based on consumer insights. If companies want to adopt this suggestion, in addition to looking at the accuracy of a bunch of models, they also need to know why such designs are more likely to win the favor of consumers. However, currently pre-trained model AI and neural network AI do not have these capabilities." During the conference, Zhou Bowen, chairman of JD Technology Committee, said in an interview with Sina Technology.

In Zhou Bowen's view, the biggest challenge facing artificial intelligence at present is still how to make AI more credible in terms of robustness, replicability, explainability, privacy protection, value alignment, and responsibility.

With the acceleration of the pace of new AI infrastructure construction, the pace of commercialization of artificial intelligence technology is indeed accelerating. However, for more AI technologies, how to ensure that they are used in a more reliable and people-oriented manner will still be something worthy of long-term attention and in-depth research.