[Explanation] If artificial intelligence also becomes a "grid", as long as you purchase access, you can enjoy intelligent services at any time. Have you thought of this kind of future?
On September 24, at the Zhongguancun Forum, talking about the prospect of artificial intelligence usage, industry experts described such a scene.
[Concurrent] Huang Tiejun, Dean of Beijing Zhiyuan Artificial Intelligence Research Institute
The future usage of artificial intelligence is a bit like everyone's use of electricity today. You are now doing various energy-related activities. We have to connect the electricity, and you can connect to the grid to buy electricity.
The future usage of artificial intelligence is similar to this way. If you need intelligence, just accept such a network of intelligence services. Your information system includes our various products, and its intelligence level is improved.
[Explanation] On the same day, a digital person for the Winter Olympics sign language broadcast based on the "Enlightenment" artificial intelligence model of Zhiyuan also made a grand appearance.
The large artificial intelligence model is a product obtained through training with the support of big data and large computing power.
Huang Tiejun said that with the blessing of large models, the intelligence level of artificial intelligence will be higher and more versatile.
[Concurrent] Huang Tiejun, Dean of Beijing Zhiyuan Artificial Intelligence Research Institute
(Large model) In a sense, it has to read all the information that can be read in the world. It uses artificial intelligence algorithms to condense this information into a representation of its internal implicit knowledge after training in computing power. Therefore, its intelligence level is higher than that of the traditional model that we say does not use so much data and does not have so much computing power training, and it has versatility, so it can meet the needs of various applications.
[Explanation] Huang Tiejun believes that large models are becoming the development trend of the artificial intelligence industry.
At the same time, there are corresponding advantages in studying large models in China.
[Concurrent] Huang Tiejun, Dean of Beijing Zhiyuan Artificial Intelligence Research Institute
The more prominent advantage of (China) is rich data resources and rich application scenarios, which in fact provides a good material condition for the generation and training of large models.
And if after the model is made, the general ability of the test model is good, and the application scenarios are abundant, then it is a good test condition.
[Explanation] The research and development and application of artificial intelligence are inseparable from the support of a large amount of data, and the safety and ethical issues of its research and development have also attracted much attention.
In Huang Tiejun's view, on the one hand, we must pay attention to these problems during research and development, and on the other hand, we cannot solve these problems overnight.
[Concurrent] Huang Tiejun, Dean of Beijing Zhiyuan Artificial Intelligence Research Institute
The ethical security risks brought by people, large models, and artificial intelligence technology are also required to untie the bell. These ethical security risks are brought about by the continuous advancement of technology.
We solve its problems mainly by relying on technical means to continuously solve it.
(Reported by reporter Shan Lu from Beijing)
Editor in charge: [Li Yuxin]