Beijing: In the sophisticated rapid scene of artificial intelligence (AI), open source models appear as a strong incentive for technological democracy and global cooperation.
“The open source AI represents more than just a technological strategy-it is a transformational approach to global innovation,” said Zhou Honggi, a member of the Chinese Political Consultative Conference (CPPCC) and founder of the 360 Chinese Security Group, in reports by the Chinese economy network on Wednesday.
The open source artificial intelligence mainly challenges traditional closed technological ecosystems. “First of all, the open source will triumph over the closed source systems,” Zhou confirms.
It highlights Deepseek, an open source AI model, as a major example of this typical shift. Within a short period, Deepseek turns from an unknown entity into the standard of industry that companies and developers worldwide are eager to adopt.
The open source approach creates a unique environmental system of cooperative innovation. “By going open source, a mechanism is created as companies and developers naturally choose to build up applications on top of the platform,” explains ZHou.
Although a company like Deepseek may not directly indicate its technology, the revenue is great: global talents, including developers, engineers, professors and doctoral students, contribute to improving technology, and creating what Zhu describes as a “biological sensation” of technological development.
One of the most persuasive aspects of open source artificial intelligence is the ability to settle the technological stadium. “For many countries that lack financial resources and technical expertise, open source models such as Deepseek provide an opportunity to develop their artificial intelligence models,” Zhou notes.
This model penetrates the national borders and technological barriers, which enhances an environmental system for open and comprehensive innovation where all countries can equally participate in developing artificial intelligence.
The approach has already shown remarkable success. The major Chinese technology companies such as Baidu and Trent and the three communications operators have combined open source models. On the international level, companies such as NVIDIA, Microsoft and Amazon, at the beginning, began to integrate these models into their ecosystems.
At the international level, China was pre -emptive from open source artificial intelligence as a model for global cooperation.
Last December, China and Zambia participated in a meeting of a group of friends for international cooperation in building capacity from artificial intelligence at the United Nations headquarters.
Representatives attended more than 80 countries and some United Nations agencies, expecting the group to enhance the cooperation building cooperation of artificial intelligence, governance, and the closure of the digital gap.
This vision extends beyond technological competition-which represents a new model for global cooperation, as technological progress is seen as a common journey instead of zero game.
Treating the critical issue of the integrity of artificial intelligence, Zhou provides an accurate perspective.
“The lack of development of artificial intelligence is the greatest insecurity,” he says. He believes that the integrity of artificial intelligence should not be an exaggeration in the transformation, but it systematically approaches.
The main challenge lies in the sensitivity of the model for manipulation, including “artificial intelligence hallucinations” (phenomenon where artificial intelligence generates wrong information), potential manipulation attacks, and access to unauthorized information.
While Zhou suggests that hallucinations can be mitigated by the online knowledge base corrections, the integration of the institution’s knowledge library, and the multi -style verification approach.
Interestingly, he looks at the “artificial intelligence hallucinations” not as purely negative features, but as an appearance of intelligence and creativity.
Zhou proposes a revolutionary approach called “model to the model”, which includes the use of smart models to manage access to the base of knowledge, control the calls of smart factors, reduce the basic model “nonsense” and attempts to manipulate.
