GPT Chat: When AI Looks Like Humans | For now, the system raises more doubts than certainties

ChatGPT came onto the public scene as a sensation at the end of 2022 and in January it surpassed 100 million active users. Artificial intelligence, developed by OpenAI (by Elon Musk), at its core is nothing new: it consists of a chatbot like any other that is used every day by private (for example, telecom companies) and public (Boti, belonging to the Buenos Aires government). Iris)). but, Its main advantage lies in its unmistakable accuracy, to the point that it is able to answer questions and provide information on a range of topics, in real time, in friendly and coherent language.; So much so that at times it seems to be imitating a human conversation. In this way, AI is valued more the less artificial it appears to be.

Chat GPT is a turning point when compared to other brute intelligence. With an almost unlimited number of pre-programmed responses, it seems to fit dialogues a little more tightly. It “seems” because, unlike promises, when the user experiences the system in his or her own body, the Generative Generative Pre-trainer Transformer chat failures are intrinsic. Countering this, the company rightly argues that a neural network will improve as users are trained on it and the responses it gives can be calibrated to specific contexts. The system takes data online; In fact, it handles so much information that some tech experts speculate that it may rival Google.

“Chat GPT has reached a peak in downloads, marking a historical milestone, but it is still not integrated into people’s daily lives. In my agency, we use it as another resource, but it still does not replace Google. What will be revolutionary is the use of artificial intelligence to generate long-term solutions. An irreversible path is suggested“, Confirms Marina Damelwho works and teaches at Pixel Digital Academy’s Digital Marketing School.

Although it seems crazy that a new technology could displace the colorful lettering giant, the fact is Google has already taken the hint and announced, through its parent company Alphabet Inc, the launch of “Bard,” its own intelligent chatbot.. At the moment, it is in the testing phase and has not met expectations. In appearance, at the beginning of 2023 Bill Gates He noted that the technology landscape will be driven by artificial intelligence, and he was only anticipating where a problem he knew firsthand would come from. ChinaFor its part, the engines are already starting and it is also joining the discussion on the applications of artificial intelligence. In fact, they have from the state communicated support with incentives of various kinds to those companies that are in a position to develop conversations similar to GPT. Competition is introduced.

See also  US Congressman arrested for forging his resume | George Santos' last lie: plead not guilty

The success of this chat in such a short time is known but not surprising. theoretically, It allows many tasks that humans would find tedious to be performed and solves within seconds. Not only does he answer questions, but he also corrects texts, translates languages, creates summaries, and solves equations.

Good and Evil

“We’re very used to using chatbots. The advance this particular one offers is that it has parameters that simulate slightly more complex conversations, based on data already available on the Internet and other data that users provide as they train them. Since the dialogues look more natural, they attract more attention“, Believes Natalia Zuazo, specialist in politics and technology, and consultant to UNESCO. In this regard, Damel points out: “It seems to me to work well but it still has some limitations, such as being a web application; connection problems also appear. I have tried it at different times and many times it does not load or does not sync well with information However, it falls short of the level of error that Bard had, Google’s solution.

ChatGPT can too Writing poems and speeches, authoring entire research papers, and publishing news stories like this one in an instant. It is a tool with great potential and like all technologies – essential to humanity – it has both positive and not-so-positive uses. While AI cannot yet replace humans, some organizations are taking precautions.

in FranceSciences Po pioneered the banning of the GPT chat bot from its students, as it considers it Copy can result in severe penalties and expulsion of a student who uses it without permission. Their powers declared that it would be different if the teacher could use them. Of course, the tool can be used to solve the cumbersome jobs related to searching for information and typing more or less drinkable in a matter of seconds.

See also  The French judiciary approved the three-year prison sentence against Nicolas Sarkozy on charges of corruption

The problem is not just copying, but copying badly. In this regard, Zuazo highlights: “When using already published information, it uses various sources on the Internet. With that said, sometimes you can stick to the truth and sometimes not. Chat is not willing to validate what you share, but to collect information and respond based on it. So, in the end, it’s up to the people to decide what’s right.”

So, there is a risk associated with honesty. In fact, a good strategy to check its effectiveness is to chat with the bot about problems in which the user is an expert, or areas of which they have advanced knowledge to check the operation of the AI. “Keep in mind that it is not a mature system. Generate answers constantly, progress is fast, but there are still struggles“, he maintains. Under this premise, from Open AI, it has already been launched GPT plus chat: A paid version, pre-calibrated better than the previous one and is already available in the US.

a Positive useon the other hand, it is he who is associated with it Coco: A psychological support tool that integrates Chat GPT, and is able to provide emotional support to people who need to protect their mental health. Its main objective is to prevent individuals in a weakened state from harming their health. According to Rob Morris, co-founder of the platform, as of January 2023, more than 4 thousand people have received Coco’s virtual assistance.

Faced with the dangers that the most horrific visions foretell, Damel takes the opportunity to highlight the positive. “It will be possible to do many things more easily and efficiently, such as writing organic content. With the advancement of digital media, many people have lost the ability to express themselves. I think AI comes in to make up for a little bit of this perceived lack of communication in HR. It even teaches how to express ourselves better, as well as how to make better use of the information available.”

See also  Saws, tattoos and soul: the carpenter who learned to make furniture for his home and succeeded on YouTube

Revolution or organization?

Artificial intelligence, sometimes surreptitiously and sometimes more outright, is changing the world. AI is very efficient at performing repetitive tasks and at analyzing a huge amount of information, although at present (and only for the time being) it does not have the ability to perform creative functions, to the extent that it proves incapable of comprehension. contexts, and does not contain metaphorical interpretation that assesses plots, subplots, surfaces, and depths. Since it feeds on data without considering its veracity, it is generally inclined to reproduce false information, and to fall victim to manipulation.

For this reason, professionals call for caution. “It is probably an exaggeration to say that chatting can be revolutionary. But it will change industries, we are sure of it,” Zuazo highlights. In this way, while it is true that many jobs will disappear, new ones will also appear. “We are on the way to replacing man in various aspects, and a new business will emerge that no longer has to do with operational issues, but with the use of reason and manipulation of tools.. On the one hand, this would be a relief, although it is also true that more complex problems, such as computer security and data manipulation, must be solved, ”comments Damel.

Meanwhile, geography is plagued by unregulated technologies that seem to evade all controls. Systems that advance so quickly, at such a pace that it sometimes seems that they have no time left to demand explanations. From the White House, Joe Biden’s government announced in October 2022 a Artificial Intelligence Bill of Rights. It is a plan that does not specify specific enforcement actions, but is presented as a call to action with the goal of protecting digital and civil rights, given the advancement of artificial intelligence. The usual sentiment remains: while technological developments advance, the subsequent understanding and organization walk on the sand and in flip flops.

[email protected]

Freddie Dawson

"Beer specialist. Award-winning tv enthusiast. Bacon ninja. Hipster-friendly web advocate. Total social media junkie. Gamer. Amateur writer. Creator."

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top