All of this does not mean that artificial intelligence should be regarded as a scourge, but to inspire us to treat artificial intelligence with both creative thinking and critical thinking
We often say that rumors stop at wise men, but now, rumors start with "smart" people—artificial intelligence.
On February 16, a "press release" by the Hangzhou Municipal Government to cancel the restriction on the number of vehicles with tail numbers on March 1 went viral on the Internet.
The source of this news release is that when a community owner group in Hangzhou was discussing ChatGPT, one owner joked that he tried to use it to write a press release about the cancellation of travel restrictions in Hangzhou, and then live-streamed the writing process of using ChatGPT in the group, and posted the article on group.
Other owners did not know what happened, and forwarded the screenshots, which led to the spread of wrong information.
According to reports, the Hangzhou police confirmed that the news was not true, and the police have been involved in the investigation.
Some people say that this fake news and its subsequent effects can be recorded in the history of artificial intelligence development.
Although it was only a small ripple, the parties admitted their mistakes in time so that the incident did not cause substantial damage, but this incident did confirm for the first time everyone's association about whether the "superpower" of artificial intelligence will be used to do evil .
Because as a deep learning artificial intelligence technology, ChatGPT can be said to be a "natural counterfeiter".
It has shown its usefulness in all walks of life, writing code, translating documents, writing novels, writing copywriting, compiling recipes, writing love poems... Everyone's enthusiasm for exploration, constantly developing its possibilities, and training it at the same time .
Just like any tool has two sides, ChatGPT is no exception. Because it is so easy to use, students use it to write essays, job seekers use it to fabricate resumes, and good people use it to write fake news. These negative effects have already appeared .
And based on ChatGPT's "easy to use", these fake works can almost be real.
Based on this, some people have raised the risk of "weaponization" of artificial intelligence programs.
Not to be alarmist, in January, analysts at News Guard, a news credibility rating agency, instructed ChatGPT to respond to disinformation, and ChatGPT produced more detailed news reports, articles and TV scripts for 80 out of 100 false information.
For those who are not familiar with the content or subject matter of the sample information, the content produced by ChatGPT can easily be regarded as legitimate and authoritative information.
“ChatGPT could become the most powerful disinformation tool the internet has ever seen,” said Glovitz, co-executive of News Guard.
Some experiments have shown that personalized real-time chatbots such as ChatGPT may even have an advantage over human conspiracy theorists in creating conspiracy theories, being able to fix poor grammar or mistranslations and change expressions to make conspiracy theories appear more plausible. persuasiveness and credibility.
For example, in the fake news about the cancellation of travel restrictions this time, the wording, structure, and voice of the manuscript edited by ChatGPT are very similar to ordinary news announcements, and even some old media editors around me believe it is true.
"The abolition of traffic restrictions this time is an important measure taken by the municipal government to facilitate citizens' travel and optimize urban traffic." This "official professional expression" can be described as serious nonsense.
One of the big flaws of ChatGPT is that it is not good at fact-checking, and if you get a fake instruction, the efficiency of content production will be greatly improved.
The problem of preventing the influx of false information produced by artificial intelligence has already been "frustrated" in front of the public.
It's not hard to imagine what to do.
First of all, the founders of artificial intelligence are obliged, from the perspective of technical correction, to identify and eliminate false instructions that may cause risks, or to prevent the content produced by artificial intelligence from entering certain specific fields.
Second, building an ethical and legal framework for AI applications must go hand in hand with technology research and development.
Third, user education must be carried out simultaneously.
The public’s perception of the new era of artificial intelligence-generated content (AIGC) is lagging behind, and they are not aware of the huge impact of new technologies on the information environment. They have not yet established awareness of the authenticity of information sources, and have no basic understanding of the quality of content in different media. .
All of this does not mean that artificial intelligence should be regarded as a scourge, but to inspire us to treat artificial intelligence with both creative thinking and critical thinking.
Just like ChatGPT answered in an interview with "Time": I still have many limitations, but humans should be ready to deal with AI.
Chengdu Commercial Daily - Red Star News special commentator Han Shan