▲Photo/Visual China
This article was first published at Southern Weekend
Text / Meiyijun
Editor / Chen Bin
In the 2023 New Year, the hottest topic in the IT industry is the new chatbot model ChatGPT released by OpenAI, Microsoft's artificial intelligence research lab, on November 30, 2022. In less than two months, its daily activity has exceeded 10 million. US digital media company BuzzFeed announced the use of ChatGPT to create content, and its stock price soared 306.94% in two days.
ChatGPT is an AI-driven natural language processing tool from OpenAI, using the Transformer neural network architecture. GPT-3.5 architecture, with language understanding and text generation capabilities, it trains models through massive databases and human language databases, making it almost omniscient, and can also understand human semantics through deep learning according to the logical relationship between the context of the chat, simulate chat communication similar to real humans, and even complete tasks such as writing emails, video scripts, copywriting, translation, and code.
ChatGPT also presupposes a certain code of ethics that says "no" to questions and requests that are judged to be ill-intentioned. Once the user's words are found to contain malicious intent, if there is criminal intent, they will refuse to provide a valid answer.
In 2019, Microsoft invested $1 billion in OpenAI in exchange for an exclusive license for its technology. Now, OpenAI's valuation has skyrocketed to $29 billion. Encouraged by this, Microsoft CEO Nadella revealed that he plans to integrate AI tools such as ChatGPT and DALL-E into Microsoft's products, including Bing search engine, Office Family Bucket (including Word, PPT, Excel, etc.), Azure cloud services, Teams chat programs, etc.
Others have proposed "ChatGPT+WebGPT": WebGPT is an advanced version of the web crawler that extracts information from the Internet to answer questions and provide corresponding sources. This feature further enables AI chat messages to be updated in real time, dubbed "'ChatGPT+' will open the door to imagination".
ChatGPT quickly became popular in the short term, shaking the AI industry boss Google, and its CEO Sundar Pichai issued a "Code Red" within the company, and claimed that for the ChatGPT challenge, it will comprehensively adjust the research and development work in the field of AI in 2023, and even invite L.E.Page, S.Brin, the two semi-retired founders, back to lead the cross-door seminar of the two major business units of AI and search.
ChatGPT shockwaves affect the world, ChinaZ.com news on January 30: Baidu's Hong Kong stock price rose 4.02% to HK$142.4, with a total market value of about HK$394.2 billion. The reason is that there are rumors that Baidu will launch an AI chatbot similar to ChatGPT in China in March 2023.
Faced with the power of ChatGPT, ordinary users cheered. But the world's top leaders in thought and technology are worried. In his tweet, Musk argued that "We are not far from dangerously strong AI."
In the AI community, ICML (International Conference on Machine Learning), one of the well-known machine learning conferences, recently announced that it is forbidden to "publish papers (except related research) containing text generated from Large Language Model (LLM) (such as ChatGPT)" to avoid "unintended consequences and unsolvable problems".
Sciences Po announced on January 27 that "students are prohibited from using ChatGPT software to complete any written assignments or presentations, except for specific course purposes and supervised by a course leader." The university believes that the move is aimed at preventing academic fraud and plagiarism, and students who use the relevant software may face expulsion penalties and may even be expelled from the French higher education system.
Some public schools in New York City and Seattle have banned ChatGPT for assignments, and some U.S. universities have announced plans to increase the weight of handwritten essays and oral exams. Holden Thorp, editor-in-chief of Science, pointed out that "ChatGPT is fun, but it is not the author", noting that "originality" is the basis of Science publications, and the act of writing text using ChatGPT is equivalent to using the software for plagiarism.
In addition to the above intellectual property and original copyright protection measures, Musk's concerns about AI over the years have been more human-level philosophical, and his concerns have proved to be not unfounded.
Just as we celebrate the Spring Festival and ChatGPT continues to be popular these days, the wave of layoffs in the US IT industry hit like a tsunami, in just ten days, Microsoft announced that it would cut 10,000 employees (its largest layoff in more than 8 years), Google's parent company Alphabet announced 12,000 layoffs (the largest layoff in its 25-year history), IBM announced 3,900 layoffs, and SAP announced 3,000 layoffs. In addition, on January 4, cloud service SaaS leader Salesforce announced that it would cut 8,000 employees, and e-commerce and cloud computing giant Amazon announced that it would cut 18,000 employees... According to Layoffs.fyi, the global technology industry laid off more than 150,000 jobs in 2022, and more than 200 technology companies have laid off more than 68,000 employees so far in 2023.
BuzzFeed, whose share price soared 306.94% in two days using ChatGPT, announced a 12% layoff (about 180 out of 1,500). Reducing payroll means higher returns for investors, but ChatGPT is just a powerful tool, and the world still runs on the will of those who hold power and wealth.
The "dangerously strong AI" that Musk is worried about has another meaning: under the impact of AI, many people will return to a state of ignorance and decadence, destroyed in chatting with AI, that is, the so-called "entertainment to death". This is caused by human nature and human physiological characteristics, human nature does not like to think and exercise, because thinking and movement will consume a lot of energy, making people feel tired and hungry, so the vast majority of people do not exercise, less exercise, like to accept simple answers. Therefore, fitness, reading, thinking, and research are the characteristics of human elites, and after stopping to forge, study, and think, elites will also sink.
For example, the United States: industrial migration began in the 1990s, and for three decades Americans were unfamiliar with industry, and the powerful industrial system that won World War II became the "Rust Belt". Then, about 220,000 IT talents who have been laid off in the past two years, after more than ten years of hard studies and personal struggles, as a result, many people can only swipe their mobile phones to chat with ChatGPT to vent the pain of unemployment, and the other side is not even a real person, just a "flower in the mirror, water moon" program. A few years later, the former IT elite became a greasy uncle, and after chatting about his savings, he could only apply for cleaning and security positions.
A friend who used to engage in research and development in a famous IT company, less than 40 years old, was laid off in 2019 due to the trend of "laying off elderly employees over 35 years old", and now he can not return to the technology industry, he said that he has been a generation behind in 4 years, and he can no longer catch up in middle age. If you know this, why study those majors? It is better to speculate in stocks from the university, this is the sinking of the elite, mourning is greater than death.
It can be seen that the shock wave brought by AI represented by ChatGPT is not only at the level of science and technology and the market, but also has a profound impact on human life patterns, social structure, and ideology.
(This article is only the author's personal views and does not represent the position of this newspaper)
