Peking University, Februray 15, 2023: ChatGPT, a chatbot developed by
OpenAI and launched in November 2022, has broken the stylized
communication of traditional AI systems and realized the transition from
providing data to solutions, said Chinese experts.
(Photo: Unsplash.com)
“First of all, ChatGPT has met huge social needs,” said Xue Lan, head of Schwarzman Scholars of Tsinghua University.
“Its
functions of chatting, answering questions and solving problems have
met the most basic communication needs of human beings. Its answers show
strong coherence and some logical ability, breaking the stylized
communication of traditional AI systems and greatly stimulating people's
curiosity,” Xue said.
Wang Hongguang, executive director of
Peking University's China Center for Strategic Studies, believes that
ChatGPT is capable of collecting and refining the information that is
searched by Google, Baidu, and other search engines, making large
amounts of fragmented information logical or even presenting it through
an article, which has realized the transition from proving data to
solutions.
“This is a great breakthrough,” said Wang.
ChatGPT
is generative. It can communicate with users constantly, generate new
content, and efficiently accomplish natural language processing tasks.
In
the “self-introduction”, ChatGPT admits it has some limitations,
including the possibility of producing wrong and biased information and
limited knowledge of the world and events after 2021.
Xue pointed
out that ChatGPT can help improve people’s work efficiency and serve in
specific fields, for example, writing documents for lawyers and
advertisement copy, and generating webs.
Some jobs and industries
such as programming and higher education will be challenged due to
their traditional thinking and skills, Xue added.
At present,
some academic journals have put restrictions on use of ChatGPT to write
or edit papers, saying AI cannot replace authors to complete key tasks
such as data interpretation and drawing conclusions.
Universities
in the U.S., France, and Australia have issued ChatGPT use bans to
prevent cheating and plagiarism. They are concerned about the threat to
independent thinking caused by too much technology.
Considering AI’s development level, it is unnecessary to set a boundary for AI right now, Wang said.
Source: ECNS
