¿La intelligencia puede separarse del cuerpo?

¿Qué relación hay entre la mente y el cuerpo?

It is feasible that the thoughts is just like the management of a online game; Desplaa el cuerpo por el mundo y lleva a paseos divertidos. O tal vez es el cuerpo el que manipula a la mente con hambre, sueño y ansiedad, algo así como un río que conduce una canoa. ¿Acaso la mente funciona como ondas electromagneticas que salen y entran a nuestros cuerpos fluorescentes? ¿O es más un car sobre una carretera? ¿Un fantasma en la máquina?

Perhaps there isn’t any metaphor that’s fully enough as a result of there isn’t any distinction between the thoughts and the physique. Only exists expertise or some sort of bodily course of, una Gestalt.

Estas preguntas, que han tormentado a los filósofos durante varios siglos, están cobrando un renovado apremio a medida que computadoras sophisticadas que cuenta con inteligenia synthetic comizan en infiltarse en la sociedad. En cierto sentido, los chatbots, como el GPT-4 de OpenAI and el Bard de Google, tienen mentes. As they’ve been educated in an unlimited quantity of human languages, they’ve realized to generate novel combos of texts, pictures and even movies. Con las indicaciones correctas pueden expresar deseos, creencias, esperanzas, purposeos y amor. Pueden hablar de introspección y duda, de confianza en sí mismos y arrepentimiento.

Sin embargo, some researchers in synthetic intelligence guarantee that know-how won’t obtain true intelligence or actual understanding of the world till it capabilities with a physique that may understand its surroundings, react to it and really feel it. Para ellos es un error y es incluso peligroso hablar de mentes inteligentes incorpóreas. An AI that can’t discover the world and know its limits, in the identical means that youngsters uncover what they will do and what they cannot, might commit deadly errors and attain their goals at the price of the welfare of human beings.

“From a quite simple perspective, the physique is the pillar of even handed and clever motion,” stated Joshua Bongard, a robotics specialist on the University of Vermont. “Según lo veo, este es el único camino hacia una intelligencia synthetic segura”.

In a laboratory in Pasadena, California, a small group of engineers has spent the previous few years growing one of many first hyperlinks between an intensive language mannequin and a physique: a turquoise robotic known as Moxie, which is concerning the dimension of a child. humano. Este robotic tiene una cabeza en forma de lagrima, unas suave manos y unos vivaces ojos verdes. Inside its inflexible plastic physique is a pc processor that executes the identical sort of software program as ChatGPT and GPT-4. Los manfacturers de Moxie, que pertenecen a una empresa emergente llamado Embodied, califican este artefacto como “el primer amigo robotic de IA del mundo”.

The robotic was conceived in 2017 to assist youngsters with developmental issues apply emotional consciousness and communication abilities. When somebody talks to Moxie, the processor converts the sound into textual content and this textual content feeds a really intensive language mannequin, which in flip generates a bodily and verbal response. For instance, Moxie can transfer her eyes to console us for the loss of life of our canine and she will be able to additionally smile to encourage us to go to highschool. The robotic additionally has sensors that obtain visible cues and reply to corporal language by imitating and studying the conduct of individuals round them.

“Es casi como una comunicación wi-fi entre los seres humanos”, commented Paolo Pirjanian, robotics professional and founding father of Embodied. “Literalmente comienzas a sentirlo en el cuerpo”. Pirjanian identified that, with time, the robotic improves in this kind of exchanges, del miso modo que un amigo que llega a conocernos mejor.

Researchers at Alphabet, Google’s dad or mum firm, have adopted an identical strategy to combine intensive language fashions into bodily artifacts. In March, the corporate introduced the great outcomes of a robotic known as PaLM-E, which might assimilate visible traits of its surroundings and details about the place of its personal physique and translate the whole lot right into a pure language. This allowed the robotic to know what a part of the house it was in with respect to different issues after which open a drawer and take out a bag of fries.

Los specialists affirm that this kind of robots can carry out primary duties with out particular programming. Podrían servirte un vaseo refresco, prepararte el almuerzo o levantarte del soilo after a nasty fall, all as a response to a sequence of quite simple orders.

Sin embargo, many researchers have doubts about the truth that the thoughts of those machines, though structured on this modular kind, is de facto related to the bodily world and, due to this fact, won’t ever have the ability to recreate basic elements of human intelligence.

Boyuan Chen, an professional in robotics at Duke University who’s engaged on the event of clever robots, identified that, decided by years of evolution, the human thoughts—or, for that matter, the thoughts of some other animal— It is inseparable from the way wherein the physique acts in the actual world and the way it reacts to it. Los bebés humanos aprenden a recoger objetos mucho antes de aprender a hablar.

In distinction, the robotic’s thoughts, with its synthetic intelligence, was constructed fully based mostly on the language and infrequently makes frequent sense errors that derive from coaching procedures. Le hace falta una connexion más profunda entre lo físico y lo theoreticalo, commented Chen. “I consider that intelligence can’t be born with out the attitude of bodily manifestations”.

Bongard, de la Universidad de Vermont, concuerda. In the final a long time, he developed small robots manufactured from frog cells, known as xenobots, which might execute primary duties and transfer round their surroundings. Although xenobots don’t look as spectacular as chatbots that may write authentic haikus, in actuality they may very well be nearer to the kind of intelligence that pursuits us.

“Intelligencia corporal no es azotar un cuerpo contra un cerebro”, assured Bongard. “Tiene que reacta contra el mundo y observar cómo el mundo reacta en consecuencia”.

Bongard additionally believes that makes an attempt to land synthetic intelligence within the bodily world are safer than different various analysis tasks.

Recently, some specialists, amongst them Pirjanian, expressed in a letter their concern about the potential of producing a synthetic intelligence that, with indifference, might surpass human beings within the pursuit of some goal (like producing paper clips. ), o aprovecharse para objective perversos (como en campañas de disinformación). The letter known as for the momentary suspension of the coaching of fashions stronger than the GPT-4.

Pirjanian identified that his personal robotic may very well be seen as a harmful know-how on this sense. “Imagina que tienes un robotic compañero de confianza que se siente parte de la familia, pero que sutilmente te está vando el cerebro”, he stated. To keep away from this, his group of engineers educated one other program to observe the conduct of Moxie and mark or stop any probably dangerous or complicated factor.

Sin embargo, cualquier tipo protección contra estos peligros será difícil de integrar en grandes fashions de lenguaje, particularly a medida que se vuelvan más poderosos. Si bien muchos, como GPT-4, están entrenados con retroalimentación humana, lo que los imbuye de ciertas limitations, el méthode no puede tener en cuenta todos los escenarios, por lo que se pueden eludir las barreras de seguridad.

Bongard, al igual que different scientists of the world, opina that the letter that requested a suspension of the analysis might generate uninformed alarmism. Sin embargo, le preocupan los peligros de nuestra tecnologia en constante crecimiento y believes that the one solution to impregnate corporal synthetic intelligence with a stable information of its personal limitations is thru the fixed trial and error technique that entails transiting by means of the actual world. .

Let’s begin with easy robots, commented “and after we reveal that we will do issues safely, we present extra legs and arms, extra instruments”.

So, maybe with the assistance of a physique, a real synthetic thoughts emerges.

Audio produced by Adrienne Hurst.

Oliver Whang es becario de reportería del Times y se emfoca en ciencia y salud. @oliverwhang21

Leave a Comment

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