artificial intelligence doesn’t Skynet, it is not a super intelligence or self-awareness. There’s even a joke that that’s all a computer can’t do: once it does, it stops being a computer to naturalize itself. What is clear is that there is no single definition. And this is problematic.
While it is true that it does not exist (yet) as in some films, what has managed to take it one step further is another of the proposals of one of the fathers of modern computing, alan turing: that of a machine that can learn, what is known as machine learning. This is what ChatGPT is all about.
“Open AI is the name of an artificial intelligence research and development laboratory with the express goal of making programs available to the general public. We try to prevent a possible existential risk produced by AI from being aggravated by the concentration of these resources in a few multinationals”, he explains to clarion Javier Blanco, PhD in Computer Science at the University of Eindhoven, the Netherlands.
This type of system, made possible by the impressive advance in computing power, constitutes a discipline that uses large volumes of already available data to build a program capable of recognize common patterns and create new data.
“ChatGPT is an example of a program capable of supporting various natural language dialogues. On the one hand, the system has been trained in this dialogue using techniques in which humans participate, both in accessing conversations to learn from and in the subsequent process of evaluating and improving interactions,” he adds.
“The enormous efficiency of these systems is somewhat surprising given how little computational sophistication they have, and is supported by large duly classified databases (by human methods or not) to learn from,” he explains.
And while there are advanced uses like discovering diseases in doctor’s offices or playing complex games like chess, there are other everyday applications that various experts have started to imagine: from writing a work email even negotiating a fine, or making up a bedtime story for the little ones, even planning.
here, what it is Chat GPT5 disruptive uses and, also, the risks that this artificial intelligence entails.
Write a work email
Many employees find it difficult to express themselves in writing. On more than one occasion, writing an email can be a problem. And more if it is to justify some Deadline late.
ChatGPT is a specialist, apparently, in giving convincing explanations and even inventing some reason to get out of trouble. It is enough to specify the conditions that the user wants.
Even if you are not satisfied with the result, you can always ask it to redo it and it will generate new content.
Come up with a bedtime story
“Tell me a story?” It is an issue that has put more than one father or mother in difficulty. What to invent to put a newborn to sleep?
ChatGPT is quite efficient and, from different entrances (orders, directions) can do a fairly respectable job.
ChatGPT in the kitchen
Several users have reported that the tool allows you to create different recipes for each dayalso advising of purchases to be made.
For those who take care of themselves on a diet, food plans can also be made (after consulting a doctor who gives indications, of course).
negotiate fines
This is perhaps one of the coolest features and is being tested outside the country. It is a plugin, a browser extension, which allows you to negotiate a fine: “The first robot lawyer”, explain its creators.
Is called “DoNotPay.com“, or “don’t pay dot comand, as reported, has several success stories.
write source code
The programmers have warned that ChatGPT is efficient by writing source code and also doing what is called “debug”, that is, to correct errors.
“You can write software, debug code, and even write in languages that even developers don’t know,” explained one expert.
Obviously all this must be taken with a grain of salt: it is a tool, and as such, prone to errors.
What real risks does this AI pose.
The collective fear that an artificial intelligence turns against its creator, that is, against the human being, is unfounded.
“The so-called ‘existential risk’ which arises from the possibility of self-reproduction and autonomisation of artificial intelligences today he has no means of subsistence. Fictions like Terminatorwhere a system becomes aware of itself and acquires destructive desires, it does not go from being to anthropomorphic look of cognitive systems whose operating principles are different and do not imply perceptual or emotional forms concomitant to human existence”.
But some problems appear in specific areas, like education: “This type of technology allows multiple possibilities of both use and evolution. there’s a concern in universities given that they can produce elaborate academic texts of a quality perhaps superior to that of university students, calling into question one of the most widespread forms of evaluation in the humanities faculties, especially in the northern hemisphere: essay writing”, warns Blanco.
“This is finally a small example of the disruptions these technologies can produce in different social and institutional practices. Variants of machine learning also cause all kinds of transformations in the most diverse fields, including sensitive areas like politics or economics,” he says.
Therefore, there are some caveats when dealing with huge volumes of data. “The use of massive data to learn about, for example, the electoral preferences or even inducing them is transforming representative democracy by displacing some traditional ways of constructing debates and redistributing power to convince large masses of the population,” he warns.
“The market stock marketwhich also exerts a disproportionate influence on local economies, today depends for its functioning on complex algorithms whose accelerated evolution makes them at the same time more incomprehensible”, states the expert.
“For example, when the Deep Blue program won a game against the world chess champion Garry Kasparov, alarms have gone off about the future of this game. Today it is known that no human gamer can beat the best current programs, in fact human players learn by studying how these programs play and their ability to play today has improved a lot thanks to this interaction “, he warns.
“Something similar is happening with the game of go. Given its combinatorial complexity, it was only recently that programs capable of beating humans could be built using, unlike Deep Blue, machine learning technologies (current chess playing programs also use machine learning).” , he adds.
Regarding the freedom of the curriculum, Blanco concludes: “I have no doubt that the best thing is to have free tools (free and open) that implement these programs, which would be good to integrate with different teaching programs (formal and informal) and the public communication of science (computational, in this case)”.
“It is the best way to productively integrate them into human practices and open possibilities for the best possible futures. There are many risks, but they will not be avoided by leaving the best technologies in the hands of a few information mega-corporations” , he closes.
SL
Source: Clarin
Linda Price is a tech expert at News Rebeat. With a deep understanding of the latest developments in the world of technology and a passion for innovation, Linda provides insightful and informative coverage of the cutting-edge advancements shaping our world.