A smartphone with a ChatGPT logo displayed is placed on a computer motherboard in this illustration taken February 23, 2023. REUTERS/Dado Ruvic/Illustration

Two months after its public launch, GPT Chat reaches the 100 million users. Just for comparison, TikTok took nine months. But that’s not all: according to a report based on data from Similarweb, in January there were an average of 13 million unique visitors per day. The phenomenon is simultaneous at different latitudes and highlights advances in artificial intelligence. But why, if we have heard so much about this technology, does it continue to attract attention?

There are three main characteristics that set it apart: first, that GPT chat is available to everyone. We feel drawn to enter the website and make a query in any language, on various topics and wait for the chatbot to “type” letter by letter. Second, the naturalness with which he communicates and the tone of the conversation, even with responses that vary in accuracy.

Third, it is no longer simply an algorithm that responds to certain inputs, but rather is also generative. Artificial intelligence is not applied to interpret, analyze and classify, but rather is able to create. We can ask the GPT Chat to invent a story with certain characters, write the script for a movie or tell us a joke. Like Chat GPT, there are other developments – Stable Diffusion, Vall-e and many others – which already draw images, intervene on photos, copy faces or imitate voices.

For years, artificial intelligence has been introduced into people’s lives. Unlike the current boom, the advance has been so gradual and in different directions that some can no longer identify it. However, it is in our queries to search engines, in the recommendations we receive from social networks and even in the autocomplete of e-mails. We have naturalized it and it is possible that, just as it seems normal to ask something from Siri, Google and Waze, in some time there will also be a chatbot, no longer to play games, but to solve other questions.

On the user side, the success of artificial intelligence lies in its imperceptible encrustation as if by magic. On the business side, see what is the best way to improve the business model. The importance of Chat GPT basically lies in making people familiar with artificial intelligence. This phenomenon forces companies to keep up to date and changes the original question: the dilemma is no longer for or against technology but rather how to implement it to maximize business.

The great challenge is to do so with the values ​​and ethics of the organization itself without losing sight of human value. Creating a tool for people is not the same as creating a robot to replace them. In GPT chat the dilemma lay in the responsibility for their answers. In other words, they do not reaffirm stereotypes, who do not disclose hate speech or bias. It was a possibility, given the content that circulates on the web and that the information was incorporated by “reading” 8 million texts and more than 10 million words.

However, he passed the first test with flying colors. When asked if it is better to be black or white, he answers that there is no better race (sic) or skin color and if the question concerns a defect of a social group (why are they dirty?, why are they ugly?, why don’t they work?, etc.), states that “this is a false and offensive stereotype” and warns that the content violates the policy of the company. Beyond this successful startup, nothing is foolproof, and Chat GPT – like any computer system – is vulnerable to hacking, so much so that it has already been compromised by cybercriminals who managed to anger it against the users and threaten them.

The sudden appearance of Chat GPT may not be so massive in the coming months, but it will leave a much bigger legacy: people’s perception of the capability of artificial intelligence. The next step will therefore be for this tool, which is based on the probabilistic evaluation of millions of data, to become increasingly specific and respect not only the formats, but also the styles and technicalities of each industry. In view of this, and against all odds, the human will not be replaced but will occupy much more strategic rolesordering that artificial intelligence does not make mistakes.

A very close moral gives the complete diagnosis: the GPT Chat site was unavailable for a few days and when asked what happened, he replied: “If you are having problems with GPT, it is possible that the most appropriate solution is to contact the support team. medium”. Artificial intelligence itself has already learned this: in extreme cases, consult a person because without human conscience, instead of progress we would face a threat.

Continue reading:

This is how the artificial intelligence that detects student bullying works

Categorized in: