We are at a time of exponential development and growth of artificial intelligence (AI) in the world. This confronts us with making decisions: get on this very high-speed train that is AI or continue at a slow pace using the tools to which we are accustomed.
We face this decision personally but it is also faced by governments, institutions and of course science. To understand how Artificial Intelligences are linked to medicine and our lives, on May 31, the Café Pesquisa: Artificial Intelligence and Medicinea space in which two experts shared their perspectives about the role of artificial intelligence in medicine and in everyone’s lives.
Luis Gabriel Moreno Sandoval, systems engineer and professor at the Faculty of Engineering of the Pontificia Universidad Javeriana; and Edwin Bautista, doctor and professor at the Universidad del Bosque, were part of this conversation.
What is artificial intelligence?
The concept of Artificial Intelligence corresponds to a set of models and technologies that seek to emulate what people do, such as face recognition, understanding language or emotions.
These technologies allow us to create devices that are capable of, for example, processing human language. That is why we are not talking about a single AI but about several, as there are several human intelligences, socially, emotionally, spatially, etc., comments Professor Luis Gabriel Moreno.
Not all artificial intelligences are ChatGPT
There are a large number of tools based on artificial intelligence, and although not all of them work like ChatGPT, they have a common element which is their efficiency: the ability to respond quickly based on a large amount of information.
When we use ChatGPT or any long language model, we feel like we are talking to someone. However, generative artificial intelligence models are the result of their training and the information with which they have been programmed.
Producing language is not the same as knowing and understanding concepts. Children, for example, learn to speak through exposure to language and as they grow they produce increasingly complex language. But that speech does not necessarily reflect a deep understanding of the topic he is responding to.
That is why it is key to keep in mind that expert and professional judgment is essential to make decisions using tools based on artificial intelligence.
Computational models
In scientific research, “a model is a reproduction of a part of the reality that surrounds us, we cannot take the lightning phenomenon to the laboratory, but we can replicate its occurrence on a smaller scale and thus be able to study it,” explained the Professor Edwin Bautista.
Artificial intelligence models work in a similar way, models of the reality that is intelligence are created and reduced to a specific task. So we have AI models that generate image, sound, text, complex calculations, and others.
Models help us represent specific problems, explained Luis Gabriel Moreno, and allow us to have tools for every need: “Just as we have tools for hammering, cutting or tearing, there are computational models with great computing capacity that allow us to solve different types of tasks.” , he claimed.
Artificial intelligence tools for medicine
Medicine also has specific problems, commented Edwin Bautista, and tools based on artificial intelligence can make the solution to these problems more efficient, for example, through diagnostic aids. Traditionally, radiologists analyze X-rays, CT scans, and MRIs one by one to identify fractures and other patterns.
With the use of artificial intelligence models, the patterns present in images can be parameterized and automated, making it possible to quickly analyze large volumes of medical images and detect injuries or diseases, following a logical process similar to that followed by a trained radiologist.
However, this does not mean that radiologists are unnecessary. They are essential for training and monitoring artificial intelligence models, as the systems are still fallible and require continuous review and adjustments. With the help of AI, radiologists will be able to devote more time to other areas that still require development and thus expand the scope of their capabilities.
With tools based on artificial intelligence, waiting times are shortened and medically relevant decisions reach patients quickly and with it, cure, care and recovery.
Build a relationship of trust
“The peace of mind with which we receive the criteria or diagnostic response from a tool based on artificial intelligence depends on the prejudices we have,” commented researcher Bautista. The trust that AI inspires is not a technical problem but a human problem, a problem that confronts us with understanding why we want the tool and what we use it for.
In medicine, trust in artificial intelligence is achieved with something “that has always been fundamental in medical practice and that is explaining medical practice to the patient in very detail,” Bautista continued. Explain the scope and limitations of AI and also of the health professional, and even the limitations of medical instruments that may lack the resolution to capture the alteration or change that is causing the disease.
Artificial intelligence models may have degrees of confidence, since, as in the human case, there may be inaccuracies and possibilities of error. We can trust what the AI says, but we cannot lose sight of the fact that they have a particular purpose. “We trust what artificial intelligence tells us, but we have to do it responsibly,” said Luis Gabriel Moreno.
Terms and conditions for artificial intelligence
The data we produce as patients, from our weight, sugar level or eating habits, are part of our medical history, but they can also be part of information conglomerates that train the functioning of artificial intelligence.
With the possibilities that this generates, risks of misuse also arise and to mediate these scenarios, the role of governments is key, says Edwin Bautista.
Legislation can establish limits, data custody criteria and protect people from selling data. Without legislation it cannot be implemented or advanced, which is why the discussion on the adoption and use of artificial intelligence is of great relevance.
The ethical use of AI is under discussion and Colombia is no exception, Moreno commented. Although the country does not have advanced technology, it does have important legal frameworks. To make the aforementioned artificial intelligence scenarios real, the Ministry of Science, Technology and Innovation should receive more investment; Colombia already has a Conpes (public policy evaluation document), which addresses artificial intelligence, as well as a committee of experts working on these issues. “We are already discussing these issues from a legal perspective,” the teacher said.
But we are not generating this technology and it is necessary as a country to make large and serious investments in CTeI, this brings a lot of progress as a country, complements the researcher.
We all happily use ChatGPT but the water is running out… where are the serious discussions we should be having about these topics?
Luis Gabriel Moreno
A necessary balance between human intelligence and artificial intelligence
There are great possibilities of finding answers by using AI. To find a balance between benefits and impacts, there are several fronts that must be addressed, commented Bautista. In addition to legislation, it is essential to educate about the social, economic, political and environmental implications of artificial intelligence. Legislation is important, but it is also crucial to recognize and address the carbon footprint generated by the use of AI.
It is key to promote the responsible and conscious use of artificial intelligence, avoiding the excessive use of resources for trivial activities and in this the academy has an important role: training new generations to understand the limitations, benefits and problems associated with intelligence artificial, in search of a sustainable balance in its use.
For Luis Gabriel Moreno, artificial intelligence is a reality that is here to stay, similar to other historical changes such as the Industrial Revolution. It is not an apocalypse, it is a new scenario where AI is a powerful tool in the hands of those who understand and use it properly. Ignoring its impact is not an option, as it is expected to replace millions of jobs in the near future and also create millions of new jobs in the process.
As humanity we have come a long way since the first uses of medicinal plants and the first algorithms of Al-Juarismi that laid the foundations of modern computing. At Café Pesquisa we also traveled an important path, we explored how these tools are integrated into our lives and how reflection goes beyond its application in medicine. Individual responsibility in the use of these tools and the citizen posture to ensure that our legislative systems prioritize the well-being of all is crucial for our role in this technological revolution to have an effect.
