The reason why AI research, which is an extension of the algorithm framework, cannot achieve true artificial intelligence is due to the lack of 'affordance'

Research on artificial intelligence (AI) is progressing rapidly, and models that draw pictures and write sentences with human-like accuracy have also appeared. However, although these are called 'artificial intelligence', in reality they are only processing algorithms based on data sets given in advance by humans, and they are true independent intelligence. It's not artificial intelligence. In order to create this true artificial intelligence, a paper has been published that 'affordance' is necessary.

Frontiers | How Organisms Come to Know the World: Fundamental Limits on Artificial General Intelligence
https://doi.org/10.3389/fevo.2021.806283

AI research began with the question ``Can machines think?'' in Alan Turing 's 1950 paper `` Computing Machines and Intelligence . I came. Nowadays, robots can automatically draw pictures and have conversations with the same accuracy as humans, but these are the results of processing by algorithms limited to a narrow range, and they are not really intelligent.

In order to deal with various things with one AI model, it is necessary to explore 'general general intelligence (AGI)', a computing system that can connect, integrate, and coordinate multiple functions. AGI can analyze, create, and act on its own, and is truly a distinguishing feature of intelligence.

AGI must autonomously set goals and improve situations without human intervention, rationally choose paths according to context and tasks, and select valuable tasks and relevant contexts from among many options. you have to choose. However, even if we try to solve tasks at a higher level, the real world, unlike the digital world, contains a great deal of incomplete, ambiguous, or contradictory information, and even humans are confused on a daily basis. .

Among them, the research team states that 'affordance' is important for AGI research. 'Affordance' is derived from the word 'to give, to provide', and is the meaning given to animals by the environment. For example, when we see an empty chair in an open space, we perceive it as something to sit on. This perception is affordance.

If you are in a dark room where you can't see anything, the first thing you do is to bump into furniture, accidentally find an electric switch, brighten the room and see where you were. This act of ``lighting the room with a switch I found by chance and confirming where I was'' is neither inference nor induction, but ``insight''. It is also insightful that mathematicians solve difficult problems from a kind of epiphany.

This insight is the human ability to deal with ambiguity and contradiction, and is the source of creativity. The research team says that what is required of AGI is to guide problem solving from interaction with the surrounding environment, and this insight requires affordance.

For example, when an AI-equipped robot picks up an object, it is practically impossible to list all the possibilities of how it can be used in advance and let it learn, and it cannot be handled algorithmically. . In other words, only living things (humans) can receive and apply the meaning of what is given. The research team argues that the development of AGI is impossible because the current AI and robotics algorithm research framework cannot identify and utilize affordances.

The research team says that biological behavior that can apply affordances is necessary for AGI, and that this discussion will have various effects not only on future AI research and evolutionary theory but also on scientific philosophy.