Revista de Marina
Última edición
Última edición

La inteligencia artificial: ¿Las llaves del edén o del averno?

La inteligencia artificial: ¿Las llaves del edén o del averno?

  • JUAN PABLO TERNICIEN NOVOA

By JUAN PABLO TERNICIEN NOVOA

  • Received at: 21/04/2023
  • Published at: 31/10/2023. Visto 574 veces.
  • Abstract (spanish):

    La inteligencia artificial ha avanzado significativamente en los últimos años; sin embargo, aún no alcanza las capacidades humanas en algunos muy complejos aspectos. La IA tiene claras fortalezas en áreas específicas, como el procesamiento de grandes cantidades de datos y la realización de tareas repetitivas con precisión y velocidad. El desarrollo de etapas más avanzadas de IA traerá grandes beneficios para la humanidad, pero tiene riesgos que deben preverse y merecen un debate sobre sus consecuencias sociales y éticas.

  • Keywords (spanish): informática, Inteligencia artificial, inteligencia artificial, automatización.
  • Abstract:

    In recent years artificial intelligence (AI) has moved forward significantly; however, it still lacks the innate understanding in complex connections that human-beings possess. AI has distinct strengths in specific areas, such as processing huge amounts of data and performing repetitious tasks with great speed and accuracy. The development of more sophisticated phases of AI will bring immense benefits to humanity, but it also has risks that must be envisage and should be debated for its social and ethical consequences.

  • Keywords: artificial intelligence, Artificial Intelligence, Artificial intelligence, automation, computing.

What do we understand by "intelligence"? In simple words, it is the ability to detach, discern, discriminate, learn, create, analyze and be curious, among other cognitive abilities that ultimately allow us to make decisions. Are these complex capabilities what we expect from artificial intelligence?

Deciding is not a trivial matter, it implies taking the initiative, making a change, affecting something or someone to create a difference, obtain a result. To decide is not sufficient just to process information, compare data and facts. When third parties are affected, that information must be analyzed under a framework of values and with empathy; that is, each situation or each creation is framed in its own circumstances, which raises the question: can we translate values, empathy and feelings into algorithms that can be inserted into these artificial brains, so that they generate ethically acceptable decisions?

Intelligence is a complex and multifaceted concept. Each person has unique abilities and strengths, where the characteristics of human intelligence are not uniformly distributed and we have tried to understand, classify, and measure them in countless investigations. Thus, some are more powerful in emotional intelligence, others in linguistics, mathematical logic, and many other classifications that, in short, highlights the individual characteristics and potentialities of each person, whose variety complements us and favors man's need to live in community. If we were to create machines with extreme capabilities that encompass all classifications of intelligence, referred to as Artificial General Intelligence (AGI), how would this alter community life? We will try to clarify what we mean by "artificial intelligence", at what stage of development we are, and, beyond the apocalyptic proposals of science fiction, what are the advantages and risks of its development at the staggering speed we can witness today.

Currently, artificial intelligence systems can perform specific tasks based on their programming and access to large amounts of data; however, these systems do not have the ability to reason, experience emotions or have a self-awareness that would allow them to act with certain independence and make decisions beyond responses established by algorithms. The artificial intelligence we use today has not yet reached the complexity and flexibility of human intelligence in terms of abstract thinking, creativity, emotional understanding, and the ability to adapt to new and shifting conditions.

This AI is known as Narrow or Weak Artificial Intelligence because it is designed to perform tasks that operate within a previously defined and limited range, there is no genuine intelligence here. It has no cognitive capabilities and even less awareness; its learning capacity is limited. Some tasks we have available today with Narrow AI are facial recognition, medical diagnostic devices, autonomous vehicles, dictation, and handwriting correction systems, among other tasks. And needless to say, the astonishing chatbots.

The real goal of many researchers is to create Artificial General Intelligence, (AGI) or sometimes referred to as Strong AI. AGI is a stage not yet reached and refers to the ability of a machine to perform any intellectual challenge that a human being may be presented with and thus understand and learn any task given to it, adapting to new and complex situations in a similar way as a person does, but with a much broader and faster access to information. AGI will be able to apply the knowledge acquired in a task to solve new and unfamiliar problems. Thus, while Narrow AI can outperform people on defined issues, AGI could outsmart humans on almost all cognitive tasks (Conn, 2023). It is this area of Artificial Intelligence that is envisioned to match human intellectual capabilities and surpass them in quantity and data processing time, circumventing limitations inherent to our imperfect human nature, such as forgetfulness, greed, fatigue, laziness, and preconceived judgments. These systems are based on the capacity for deep learning and the development of neural networks, which allow them to access and process enormous amounts of information, learn from them, and make self-determining decisions.

The fast-paced development of machine intelligence can be seen in systems with amazing features that are already available to us. Chatbots and virtual assistants can perform automated tasks and commands without human involvement, as well as accept and respond to questions in normal human language by holding a pseudo-conversation. Also, the reliability of autonomous vehicles that can navigate in complex traffic situations, choosing the best routes based on different criteria, is fast-growing.

However, it is important to note that replication of emotions and autonomy in AI systems are the most complex factors to insert in machines that do not yet reach the level of human intelligence and reasoning capacity. There are still significant technical and ethical challenges that need to be addressed to achieve a sustainable and responsible development of these technologies.

How the development of these intelligent systems will affect society is something that worries scientists and politicians as well. As proof of how sensitive this transfer of rationality between man and machine can be, in the midst of the age of knowledge and the technological development, recently the news-media published an open letter from a group of scientists and businessmen in which they have requested a 6-month moratorium on the development of artificial intelligence in order to generate a debate on its repercussions and establish certain limits (Deutsche Welle, 2023). If technology entrepreneurs such as Elon Musk and Steve Wozniak, among many others, believe it is necessary to establish limits, then we should probably pay attention to what this group of experts is concerned about.

Indeed, it is still difficult to envisage whether the development of advanced AGI systems that can replicate self-awareness and emotions would be good or bad for humanity because, like any human creation, it depends on how such systems are used and controlled. On the one hand, machines that can recognize reality, generate emotions, and learn, could have a positive impact on society by generating informed, efficient, objective, and ethical decisions in the solution of disputes and complex problems, improving efficiency in a variety of fields, such as medicine, infrastructure, administration, robotics, and space exploration, among others. Some of the most significant benefits of advanced artificial intelligence or AGI can be:

  • Productive efficiency: AGI can automate and manage a wide range of complex tasks, from the production of goods to medical care. This can diminish error and accident rates, increase efficiency and distribution, and reduce costs.

  • Security: AGI systems can contribute to security in areas such as cybersecurity, surveillance and control, crime prevention, identification of dangerous persons, weapons and explosives, marine, land, and air traffic control, among many others.

  • Health: AGI can facilitate the diagnosis of diseases and the design and control of more effective individualized treatments for different pathologies. Its participation in the management of pandemics would facilitate containment of diseases. In this sense, it is expected the implementation of better telemedicine and patient monitoring systems, incrementing the physician’s competency, and expanding the coverage of health networks in remote areas, with an empathetic approach with patient care.

  • Innovation and technological development: AGI can stimulate innovation in different areas, from energy to biotechnology, which is especially important in the face of climate change and pollution, and the mitigation of its consequences. Improving aspects such as simulation will achieve, among other effects, the optimization of the production and use of renewable energy resources, the development of new non-polluting materials, and efficient recycling processes.

  • Everyday life: AGI can improve people's quality of life, with greater access to information, recreation, or solutions to daily living challenges, such as transportation, food, or hygiene. For example, these systems could help elderly or disabled people to live independently in their homes by monitoring their health and controlling and managing their home environment. Intelligent systems with the ability to replicate emotions and empathy can help mitigate the loneliness that an increasing number of people suffer in today's fast-paced modern society.

But not everything expected from AGI is good news, some of its features may have negative social impacts, at high risk that advanced artificial intelligence systems may be used irresponsibly, maliciously, or that their control systems may be overwhelmed.

Beyond what movies and science fiction have shown us regarding the threat that the development of autonomous machines with the capacity to make decisions could imply, there are potential risks that, in the author's opinion, have not been sufficiently debated nor have there been firewalls to cancel possible adverse effects generated by the autonomous machines themselves or by their makers. Some of these negative effects may be:


  • Unemployment: AGI and its ability to control multiple processes may replace jobs that traditionally required human intervention, increasing unemployment and economic inequality. While this is an effect that other stages of production-mechanization have already experienced and that the system was able to absorb, AGI could take automation to unprecedented levels, encompassing all or most of the power-structures of business and management.

  • Privacy and security: There is a vast amount of information about each-and-everyone scattered across the network in the form of medical records, bank accounts, personnel evaluations, retail purchases, recreation, etc. AGI will have the ability to compile and process these large amounts of personal data, which would put individuals' privacy at risk. AGI can use all this information and network-management to develop more sophisticated and invulnerable cybercrime tools.

  • Autonomous weapon systems: The use of AGI in military affairs, in weapons with the ability to select their targets without human intervention, and based on a probable nebulous effect, may lead to operations outside ethical and legal boundaries, with serious humanitarian consequences. The debate will arise as to who should take responsibility when one of these weapons makes the wrong decision.

  • Control and governance: The great capability of information processing that AGI will have at the individual level in real time, will facilitate the development of tools to control crowds that can threaten essential values, such as freedom and justice, with adverse effects on democracy and its institutions. Thus, AGI will have the power to act on the values and beliefs of each person, altering our perception of reality with disinformation that, among other effects, can affect electoral processes, limiting our options.

As proof of the potentialities and risks of this technology, China has reported that it will implement regulations to the contents presented by these systems so that they reflect "fundamental socialist values" (Prieto, 2023). This presents us with a dilemma: on the one hand, applying regulations, as proposed by the Chinese government, can transform an information system into a weapon of propaganda and indoctrination. On the other hand, opening the door to unlimited development of AI, without ethical considerations, may generate risks yet undetermined. It is important to acknowledge these risks and work towards a consensus aimed at mitigating them, seeking to ensure that AGI is used ethically and responsibly for the benefit of humanity.

Most likely, “Terminator” and "Rise of the Machines"-style movie predictions will never come to completion, at least not in an apocalyptic form. But when we talk about machines capable of accurately remembering an enormous amount of data, processing it, and solving complex problems in a time and volume that far exceeds any human being, this will inevitably have profound social consequences, even altering the way we currently understand community life, especially if this decision-making capacity is not coupled by values that force empathetic and ethical behavior.

It's important to clearly understand the difference between artificial intelligence and automation. It is much more than that and its effects are finally generating a debate that is heating up. In the First Industrial Revolution, machines absorbed heavy work and multiplied it, generating great social changes and a turning point in human history. The effect we can expect when machines take over intellectual work and multiply it will undoubtedly be astonishing, creating a new turning point in the human timeline. This is a technological revolution whose effects will far surpass all others that humanity has faced throughout its history. It is now that we must discuss the controls and limits so that this technology enhances our lives and does not become future threats to our freedom.



Inicie sesión con su cuenta de suscriptor para comentar.-

Comentarios

Related Articles

Inteligencia artificial en la detección de cocaína en Chile
ESCENARIOS DE ACTUALIDAD

Inteligencia artificial en la detección de cocaína en Chile

La estadística actual indica que los niveles de narcotráfico vía marítima han aumentado en Chile en los últimos años (Mi...

Aplicación de la inteligencia artificial a la defensa y seguridad
TEMA DE PORTADA

Aplicación de la inteligencia artificial a la defensa y seguridad

“El crimen organizado se combate como región o no se combate (Tohá, 2024)1" Miles de millones de personas utilizan...

Inteligencia artificial y guerras futuras
TEMA DE PORTADA

Inteligencia artificial y guerras futuras

“La inteligencia artificial es el futuro no solo para Rusia, sino para toda la humanidad. Trae consigo enormes oportunid...

Inteligencia Artificial en la Educación Naval: Posibles ventajas y desventajas
ACTUALIDAD

Inteligencia Artificial en la Educación Naval: Posibles ventajas y desventajas

NOTA del Autor: Este articulo ha sido completamente escrito por una aplicación de Inteligencia Artificial (chatGPT). Se ...

La trinidad de Clausewitz y la inteligencia artificial
ESCENARIOS DE ACTUALIDAD

La trinidad de Clausewitz y la inteligencia artificial

La llamada “Industria 4.0” da el inicio a la denominada “cuarta revolución industrial”, donde la característica principa...