By Ada Wang

ChatGPT， a super language model that got famous overnight， has caused a stir in the quiet artificial intelligence sector. Recently， Google announced an artificial intelligence chatbot technology called Bard， in response to ChatGPT. Baidu will launch its ChatGPT-style service， Microsoft has planned to fuse ChatGPT-like technology into its own product services， and BuzzFeed has“hired” ChatGPT to write articles. After being on the market for just two months， ChatGPT has attracted the world attention.

The popularity of ChatGPT signals the huge progress of AI technology， and is likely to become a milestone in the development of the AI industry. But some experts have said that the commercialization of ChatGPT will face three issues， mentioning copyright， information technology， moral principles and supervision .

Robot that is able to chat

If you have an AI secretary， what do you want it to do for you？ Writing code， translating novels， enriching resumes， polishing articles， and even taking exams？ Today， the AI secretary named ChatGPT has been able to complete most of the tasks that you could think of.

Last November， Open AI launched an AI chatbot ChatGPT， which is able to make simple dialogue with users， and even chat with users according to the language context. It is also able to write emails， poems， code and solve math problems. ChatGPT quickly became popular across the internet and hit more than 100 registered users within five days of being launched.

Compared with ordinary search engines， ChatGPT can better understand the fuzziness of human language and answer personalized questions. For example， a user asked ChatGPT the following questions： If you were Socrates and I was your student， which aspects of me would you care about？ If you were a merchant of Venice described by Shakespeare， what would you think of me as a student who cant make money？ And ChatGPT was able to seize the different characteristics of Socrates and Venetian merchants and responded， “In Socrates view， making money is not the main purpose of life”， “Shakespeares Venetian merchants pay more attention to money and entertainment”， showing a high emotional quotient.

It has been raised that some student may use ChatGPT as a perfect tool for cheating on their academics. It only takes a few seconds to type the key description of the topic in the search box， and an academic paper will come out within several minutes.

In addition to paper writing， ChatGPT can also find bugs in programmers code. Some developers said in the trial that ChatGPT provided very detailed solutions to their technical problems， solutions which were more reliable than other types of search software they had tried before.

The high popularity of ChatGPT reflects how much AIGC （AI Generated Content） is thriving in the world. According to statistics， there are about 250 start-ups in the field of concepts similar to ChatGPT. Among the top 10 technological trends in 2023 announced by Alibaba DAMO Academy， “generative AI” is one of the main trends.

In fact， ordinary people are not unfamiliar with the field of AIGC. In the past year， “AI painting” and “AI filter”that have frequently appeared in social platforms are one of the many application scenarios of AIGC. Although it has been developing for some time， AIGC has not achieved commercialization at the terminal of end users， as of yet. The emergence of ChatGPT will change this situation.

The popularity of ChatGPT also gave birth to various small businesses providing peripheral services. A merchant engaged in account business admitted that he had earned hundreds of thousands of yuan monthly by selling ChatGPT accounts.

Arouse Enthusiasm for AI business development

The sudden popularity of ChatGPT has raised the enthusiasm of the industry towards AI business development.

According to a research report issued by UBS， ChatGPT has hit about 100 million active users within two months after being launched and has become the consumption app with the fastest growing in number of users. Many technology experts said that the emergence of ChatGPT and the launch of the paid version proved that AI technology has entered the phase of fast commercialization， which has huge potential and high productivity.

The U.S. Fortune magazine pointed out that no chat robot can match the power of ChatGPT. ChatGPT can carry out long and free dialogues， answer questions， and write various materials upon users requirements， such as business plans， poems， computer codes， and scripts. However， ChatGPT is not perfect： its results are not accurate enough； the information source cannot be quoted； the wording needs further refinement.

In January this year， Microsoft announced that it would expand cooperation with OpenAI， by investing USD 10 billion into OpenAI. The U.S. magazine Forbes commented that although OpenAI has no clear profiting model for ChatGPT， the artificial intelligence has huge development potential.

ChatGPT led the rise of technology stocks. The stock of artificial intelligence manufacturer C3.ai has risen more than 133% so far this year； small-cap stocks related to AI have also been rising. The stock price of BigBear.ai， which provides services to health care， government and manufacturing industries， has risen by more than 610% this year.

The industry pays high attention to the challenges posed by ChatGPT. The data shows that the number of robots deployed by Amazon is increasing rapidly， with an increase of about 1，000 per day. In addition， the Wall Street Journal reported that Meta， the parent company of Facebook， is also planing to invest an additional USD 4 billion to USD 5 billion in the data center in 2023， which is expected to be fully used in developing artificial intelligence.

IBM CEO Arvind Krishna said recently that the AI technology is expected to contribute USD 16 trillion to the global economy by 2030.

Technological ethics issues are still pending

“AIGC （AI Generated Content） has proved that after the manufacturing industry， the content industry is also expected to enter a new era of humancomputer collaboration.” Li Di， CEO of Xiaobing Company， believes that people should regard AIGC as a new tool to stimulate creators creativity， so that human creators can further return to the essence of content creation， namely， creative thinking. “Questions are more important than answers. For human beings， asking questions creatively will be far more important than providing answers in the future，” said He Baohong， Director of the Institute of Cloud Computing and Big Data， China Academy of Information and Communications Technology.

Although ChatGPT shows people the great power of AI pre-training model， there are still some important issues pending.

First of all， in terms of knowledge issues， ChatGPT sometimes creates unknowingly mistakes in text which are hard for non-professional personnel to detect. For example， someone searched the references of the papers generated by ChatGPT， and the results showed that the references provided by ChatGPT were actually fabricated.

“At present， the accuracy rate of knowledge obtained from ChatGPT is less than 90%.” Li Di believes that in the next five years， the best way for people to obtain knowledge and information from public channels will still be using search engines.

In addition， the technical ethics issues associated with ChatGPT have also attracted wide attention. Zhang Xiaorong， President of the Deep Science and Technology Research Institute， said that assuming that ChatGPT has the conditions for large-scale commercialization， at least three major issues need to be addressed： copyright issues， information security issues， ethics and regulatory issues. Chen Jia， an independent international strategy researcher， took the scientific research field as an example to explain his concern about the ethical boundaries of technology： “with the exponential growth of ChatGPTs analytical capabilities， it will be difficult to avoid the author from using it to forge experimental data and tamper with experimental results in the future.”

Chinese version of ChatGPT is expected to appear soon

What social effects will ChatGPT produce？ Liu Zhiyuan， associate professor of the Department of Computer Science and Technology of Tsinghua University， believes that the artificial intelligence pre-training model technology used by ChatGPT can liberate people from mechanical labor， and the Chinese version of ChatGPT is expected to appear in a year or two.

“This type of AI products is actually analyzing the keywords of our instructions when chatting with us， and then calling the corresponding task module to respond.” Liu Zhiyuan explained that ChatGPT generated by the pre-training language model is more like a real communicator. Through independent learning of the huge scale of data on the internet， it will think. It will be able to understand the needs， and also complete complex tasks， which is revolutionary progress and a landmark product in the field.

However， some people are worried that ChatGPT will be abused for writing papers， cheating in exams and other immoral uses. As for relevant disputes， Liu believes that it is inevitable that the exploration and application process of new technologies will cause conflicts with society， but it should be viewed positively from a long-term perspective.

When will the Chinese version of ChatGPT appear？ Liu Zhiyuan said that based on ChatGPTs current visible technology， Chinas pre-training model can catch up with its level in about one to two years. “The gap is not huge， but what we need is to become pioneer of the field， rather than follower. It will still take a long time for us to develop new technologies and products in the filed of artificial intelligence.”