ChatGPT: Revolution or Danger?

ChatGPT: Revolution or Danger?
Photo by Jonathan Kemper / Unsplash

On November 30, 2022, the OpenAI company presented the ChatGPT chatbot to the general public. It is a neural network with an advanced language model that can actively participate in discussions and generate extended answers to questions, analyzing a deep layer of information, predicting, taking into account previous conversations, etc., like a person. So, can ChatGPT replace human intelligence?

ChatGPT is a real revolutionary milestone in the development of artificial intelligence. It is too early to talk about a real analog of “human” intelligence, but this is a massive step on the way to creating its digital equivalent.

When asked, ChatGPT describes itself like this:

“I work with deep learning algorithms and have been trained on a large dataset of text from the Internet. By asking for a prompt, I generate a text response that is intended to be similar to what a human would respond... I can be used for a wide variety of tasks, including text generation, question answering, language translation, and more. It could impact the future of humanity by improving performance in several different industries, such as marketing, customer support, medicine, and more. However, it should also be taken into account that the use of technology may lead to the loss of jobs for some people, so it is important to continue to monitor and manage the impact of technology on the future of humanity.”

What ChatGPT does not do (yet)

It does not analyze information about current events. ChatGPT graduated in 2021 and is not intended to be used online to find new information.

It has limited access to potentially harmful information. For example, it will not be possible to get a response about information that can harm people’s lives, like the development of a campaign that contains hate speech, etc.

It does not distinguish truth from lies and can use fakes. There is a saying in IT that such models “hallucinate”: they create a certain picture of reality based on the data received, which is not necessarily true.

It will answer serious questions but advise consulting a specialist (financial advisor, doctor, lawyer, etc.).

It does not have independent critical thinking and it is precisely what distinguishes artificial intelligence from humans.

Not available for everyone, in particular, ChatGPT does not officially work in Ukraine. OpenAI explained that they did this “not to violate world sanctions due to the annexation of ORDLO (SDDLR) and Crimea in 2014.” The creators allegedly do not know how to distinguish these territories from customers from the rest of the territory of Ukraine.

The main threats to this technology

Spreading misinformation

The main problem is that ChatGPT uses all the information published and available on the Internet – without distinguishing between true and false or inaccurate information. Because of this, a chatbot can be both a source and a means of spreading false information. ChatGPT, for example, can generate the results of non-existent research, scientific discoveries, etc.

Unforeseen consequences of use

In particular, the negative attitude towards ChatGPT is in scientific and educational circles. For example, schools and academics in New York have already banned the use of this chatbot’s content. The city’s Department of Education blocked access to the AI chatbot on its devices and networks. The ban is due to a potential “negative impact on student learning, as well as concerns about the safety and accuracy of the content.”

And during the International Conference on Machine Learning (ICML), ChatGPT was called an “exciting” development accompanied, however, by “unforeseen” consequences. The first question that arises, in this case, is: who owns the result of the work of such systems? They train on public data, usually collected without consent, and sometimes report the output verbatim.

Furthermore, it is unclear whether AI-generated text and images should be considered “new or merely derivative of existing works.” It is not clear to whom to attribute the authorship – to the machine or to the person who controls it.

Can take away people’s jobs

ChatGPT may interview and be hired as an entry-level software engineer at Google, according to CNBC, citing internal Google technology tests. ChatGPT can even write simple smartphone apps like a calculator that will work, albeit with a somewhat primitive design.

Amazon employees who tested ChatGPT found it to be very good at answering customer service questions, great at creating educational documents, and very effective at answering questions about company strategy.

ChatGPT attracts attention with its content creation capabilities, such as essays and emails. ChatGPT answers have aced college-level exams, aced job interviews, and even made some schools rethink their approach to homework.

In his article for The Guardian, journalist Henry Williams says that the AI-generated material required minor changes. After editing, however, it was a text he would have paid £500 ($615) to write. Williams says the ChatGPT marketing article had a simple structure and was written in a slightly “inhuman” tone, but “got all the key points, was correct in terms of grammar and syntax”. And most importantly, it took only 30 seconds for the chatbot to write it!!

Access to ChatGPT is currently free, and the OpenAI company which developed the bot has not released official applications. Therefore, all mobile applications in the App Store and Google Play that masquerade as ChatGPT and offer some “advanced features” or some “subscriptions” are trying to make money from you.

BUT

Very soon, ChatGPT will cost $20 per month – this was announced by OpenAI the other day. This subscription will initially be open to users in the United States. According to the developers, this will provide them with faster response times and priority access to new features.

RECORD

As of January of this year, the ChatGPT chatbot had 100 million active users – and that’s just two months after its launch! According to a UBS report, this fact makes it the fastest-growing app in history.

TikTok took about nine months after its global launch to reach 100 million users, while Instagram took 2.5 years.

By the way, already in the first month of its existence (December 2022), ChatGPT had 57 million active users. By January, it was visited by about 13 million individual users per day, more than double the number in December. “In 20 years of studying the web space, we cannot think of another application that has evolved faster,” UBS analysts say.