Robin Li: AI cannot take away human jobs

In the past six months, generative artificial intelligence (AIGC) has become popular, once again arousing widespread global attention to the development of artificial intelligence.

Some experts pointed out that the next 5-10 years will be a critical period for the development of artificial intelligence. It is predicted that the scale of China's core artificial intelligence industry will exceed 1 trillion yuan in 2030, and the global artificial intelligence market will reach 16 trillion U.S. dollars. There is still great potential and space for development.

But at the same time, the powerful capabilities demonstrated by large artificial intelligence models such as ChatGPT have also caused many people to worry that large AI models will replace human work. What do practitioners in the artificial intelligence industry think of this?

On May 18, Baidu CEO Robin Li said in a speech at the Seventh World Intelligence Conference that artificial intelligence will not reduce job opportunities, but will bring about the next growth miracle of the global economy. Whenever a technological revolution occurs, some jobs are eliminated, but more new jobs are created.

He gave the example of the automobile industry, explaining that cars replaced horse-drawn carriages, causing the coachman's job to disappear. Today, the automotive industry has become one of the largest industries in the world, creating hundreds of millions of jobs. In China alone, 30 million people are engaged in automobile-related jobs.

Robin Li said frankly: "I'm not worried that big models will make jobs disappear. I think more jobs will be created in the future."

360 founder Zhou Hongyi is also optimistic about the future of large models. At the same conference, he predicted that in the future China would not have just one large-scale model, but rather each city and government department would have its own proprietary large-scale model. Not only can it provide support for leaders and government decision-making, but it also has unique large-scale models for cities and governments to assist office workers in their work.

According to Zhou Hongyi’s view, the exponential leap brought about by artificial intelligence will promote the development of intelligence, prompt the peak of digitalization to arrive driven by large models, and bring huge development opportunities. He said that the greatest significance of GPT is that humans have recoded and stored all knowledge for the first time, which represents the arrival of the era of super artificial intelligence and will have a profound impact on the development of human society.

But how do large models change artificial intelligence? What is the difference between current artificial intelligence and the past?

Robin Li pointed out at the meeting that big data, big computing power, and big models have led to the emergence of intelligence. At the same time, artificial intelligence has undergone a directional change, from discriminative AI to generative AI. Big models also redefine human-computer interaction, redefining marketing and customer service. Any company that has the best way to communicate with customers will have customers.

He said that the reason why artificial intelligence has undergone such big changes is not only the change of application scenarios, but also the technology stack behind it has also undergone very fundamental changes. For a long time, the IT technology stack that everyone is familiar with is the chip layer, operating system and application layer. Today, the arrival of the artificial intelligence era has changed this pattern.

"The current IT technology stack has become four layers. The bottom layer is still the chip layer, but the main chip is no longer the CPU, but represented by the GPU, a new generation of chips suitable for parallel large-scale floating point operations. Above We call it the framework layer, which is the framework of deep learning, such as Baidu's PaddlePaddle, Meta's PyTorch, and Google's TensorFlow. The layer above is the model layer, such as today's ChatGPT, Wen Xinyiyan, etc. , these belong to the model layer, and native applications in the AI ​​era in the future will be developed based on large models." Robin Li said.

At this conference, domestic companies such as 360 and Baidu demonstrated and introduced the latest progress of their own large models.

Zhou Hongyi demonstrated two large-scale products, including a product called "360 Intelligent Brain" and an artificial intelligence image processing tool called "360 Hongtu". He said that 360 has achieved some results in core technology, scene layout and product polishing through the "two wings flying together" artificial intelligence development strategy, but large language models require high-quality training and feedback to become more intelligent.

According to Robin Li, Baidu has deployed Kunlun core in the chip field and has launched two generations of products with a deployment volume of tens of thousands. Kunlun Core’s third-generation chip will be available early next year. We develop our own chips because it’s too expensive to buy other people’s chips. ”

He also introduced that Baidu has Fei Paddle in the framework layer, Wenxin Big Model in the model layer, and Baidu Search in the application layer. “Currently, there are more than 2 million users waiting in line for testing, and there are also hundreds of thousands of them. Enterprises hope to access Wen Xinyiyan for testing."

According to Robin Li, the greatest unsustainability and danger are not caused by the uncertainty brought about by innovation. On the contrary, stopping innovation, not inventing, creating, or making progress, and continuing according to inertia will bring about various unpredictable risks, which are the greatest threats to mankind.

Zhou Hongyi believes that large models will bring about a new industrial revolution and can be like a "power plant", processing big data that was previously difficult to use directly into "electricity" and empowering thousands of industries. He emphasized that China must catch up with the development level of large-scale models, because this is directly related to the country’s productivity level. He has no doubt that China can build its own big model. At the same time, he also reminded China to adhere to long-termism in developing its big model and to make up for its shortcomings in a down-to-earth manner.

The above is the detailed content of Robin Li: AI cannot take away human jobs. For more information, please follow other related articles on the PHP Chinese website!