Chat GPT-3 would not be controversial if it were not, before that, dazzling”, Santiago Bellomo, dean of the School of Education of the Austral University, summed up to PERFIL. Bellomo explains that “before the appearance of this type of tools, one could find the information necessary to build, for example, a newspaper article on the web. But the task of searching for sources, selecting information, building the argument, will remain the responsibility of the journalist. Today Chat GPT-3 or other available models allow the construction of the newspaper article itself, giving rise to automatic journalism.
So if you want to understand what this technology is that is mentioned in different publications and media, the dean of the Austral University details that “it is an application of the Natural Language Processors (NPL) family that uses Deep Learning from the construction of artificial neural networks. This allows the machine to simulate an oral or written conversation and build responses that respond to specific instructions. “Chat GPT-3 –he adds– works in a written format, like a chatbot. Tools like Siri or Alexa use similar technology for oral dialogues.
Perhaps among the characteristics of this type of artificial intelligence (AI) stands out the fact that it can draw conclusions. For Ariel Hache Mempert, executive director of TEDxRíodelaPlata and an expert in educational experiences, the main characteristic is that he has the ability to interact in natural language and when faced with a question as a human would in his own language. So, that makes it really very easy to use, the access barrier is absolutely minimal, with which it is something that can be used tremendously massively, explains Hache Mempert to PROFILE.
Where to apply it Dean Bellomo believes that this type of technology will surely be applied in all areas of work, especially those that are based on prediction. “Medicine and diagnosis, for example, will be easier, something in which artificial intelligence will surely be able to be used faster.”
It is not being used in education yet, but Bellomo acknowledges that technology has the particularity that young people are more likely to consume AI and if students use it, teachers have to apply it. “This tool has the potential to help that teacher to make this follow-up much more personalized since it does not depend entirely on that same teacher, he already answers questions more immediately,” he highlights.
At the same time, Bellomo acknowledges that Chat GPT “is tremendously powerful, as well as disruptive because you have to change a lot of the things we did before because now with this type of technology they no longer make sense.”
Hello good. Faced with these new challenges for the different communities, both Bellomo and Hache Mempert highlight some points in favor of ChatGP. Among the benefits, Bellomo mentions the “undeniable” contribution of NLPs in general, and of the GPT-3 Chat, to simplify and facilitate the tasks that have to do with the search for information and the development of content. Also in the reduction of time that previously required more time and effort and are now resolved almost immediately and with very high levels of credibility, but at the same time the dean clarifies that the ease of production that he has in any field, does not however, service is accompanied by quality.
This tool also operates as a translator, which reduces language barriers and allows access to more information and services for many more people in the world, as well as produces a text in their own language and translates it automatically.
To these benefits or opportunities that ChatGPT offers, the executive director of TEDxRíodelaPlata highlights that it is a technology that has the capacity to increase the impact of what is done. “Anyone who knows how to use this well is going to have a really close ally to enhance their own abilities, even to learn on their own or to complement what they do,” he explains. At the same time, he considers that this tool would allow services that are exclusive to reach more people individually who could not pay for them before. “For example, access to medicine. I believe that costs will drop a lot, because if you have an artificial intelligence that is prepared to care for patients or at least to demonstrate patients, access to basic public services will improve significantly. And I say medicine, but I could say mental health or issues related to security, ”he argues.
Bad. Regarding the analysis of this type of tools, the dean of the Austral University mentions some points that concern him, especially in the educational field. Bellomo wonders if a machine is going to solve a translation, what is the need to learn a language? This applies to language learning, to the acquisition of writing skills, selection and interpretation of texts, and many other operations that these machines can solve. This is not a new problem in education. When calculators arose, this debate arose among mathematics teachers.
Another issue is that these technologies invent answers when they don’t know something, or provide poor solutions. Contrary to what many think, they do not do it on purpose. They are programmed to search among hundreds of millions of records for the information related to what one requests, and using predictive algorithms to generate an answer that meets the expectations. If the slogan is not clear, if the basic information is not there or is not enough, or if it is biased, these defects will be transferred in the response, he details.
Bellomo considers another point to work on and in a deeper way is how will teachers be able to distinguish between a production made by the students and one made through artificial intelligence? “Immediately, teachers will have to sharpen their ingenuity to the maximum to produce slogans that this type of tool cannot answer, or to have contexts in which copying or fraud are not so common or simple.” The dean raises a major question chat GPT-3 scandalizes because, in some way, it highlights the crisis of meaning that education is going through. People who prefer to simulate their own development through fraud or copying cheat solitaire and are the first victims of their self-deception.
Another issue to take into account is the inequality gap. Hache Mempert exemplifies: we know that, for example, distributing technological devices, such as notebooks or tablets, is very positive in terms of the gap in access to technological devices, but to the extent that this is not accompanied by good education around that, education, pedagogy, how to use, that increases the use gap. Because? Because there are more low-income people with little access to education who use those same tools, from technologies for leisure or for tasks that are not productive, and there are more people with more resources, better trained, using them to find work and develop professionally.
That the potential of this type of tool is still not fully understood, it can be positive or negative, which is to be expected if technology allows it, because if one thing is known it is that “the melons settle when walking ”.
Learning to live with AI
In the last decade we have been witnesses, victims and protagonists of unprecedented technological developments, such as the incredible and exponential development of artificial intelligence platforms capable of generating text with a level of fluidity, coherence and meaning that previously would be thought possible only in human beings ( and it was what ended up differentiating us in various tasks). But, as with any technological tool, the adoption of applications like ChatGPT raises many concerns, dilemmas and risks that we must take into account as a society.
The spread of deepfakes, fake news and propaganda elements is a major risk that accompanies the use of new technologies like ChatGPT (and more so in election years). Artificial intelligence (AI) algorithms have biases and some errors (yes, they are not perfect) in the training used to train them, which could lead to the dissemination of inaccurate or somewhat misleading data.
Despite the dangers and debates that come with it, technology must be recognized for its enormous potential for good. ChatGPT’s ability to generate content fluently has applications in a myriad of fields, including education and scientific research. In addition, without a doubt it helps in the automation of operational and repetitive tasks, in a theory that promises to free us to focus on tasks that require a human component, such as creative problem solving and the analysis of complex situations that not only have hard and quantitative variables.
The adoption of emerging technologies must inevitably be approached critically and introspectively. This involves understanding about the constraints and opportunities, and making an effort to reduce the negatives while maximizing the positives. One way to do this is to ensure that technologies are developed and used ethically and responsibly (to the best of our ability). Ensuring that adequate measures are taken to protect data privacy and security and that algorithms are designed to minimize bias and harm. Like everything in life, nothing in excess is good.
It is also crucial to foster a culture of digital and data literacy. This means teaching people the ins and outs of technologies so they can use them critically. Skills like critically evaluating information online and understanding how algorithms can affect the information presented to us are part of this.
I personally believe that despite significant challenges and limitations, the use of technologies like ChatGPT also offers great potential for good. The key to maximizing the benefits and minimizing the risks is that we approach technology critically and thoughtfully, and foster a culture of digital and data literacy. By doing so, we can gain the insights needed to effectively coexist with AI technologies in the future.
* Author and disseminator. Emerging technologies specialist.
You may also like