ADS
Artificial intelligence (AI) has made significant strides in recent years, and it has now surpassed a key human ability for the first time. The ability to learn the meaning of a word and apply that meaning to other linguistic concepts is a universal skill possessed by humans. We can abstract concepts and recognize objects based on their shapes, regardless of their color or composition. We can even analyze cloud formations and derive meaning from them.
This particular facet of human intelligence is known as composite generalization. Cognitive scientists, such as Jerry Fodor and Zenon Pylyshyn, have argued that artificial neural networks could also develop this capability. However, there has been little progress in this area since the mid-1980s.
Researchers from the University of New York and Pompeu Fabra University in Spain have been working on this problem for some time now and have developed a new technique called “Meta-aprendizado para composicionalidade” (MLC), which aims to replicate this specific human skill.
Their study, published in the renowned scientific journal Nature, demonstrates how AI tools like ChatGPT can successfully make composite generalizations using MLC. Field tests conducted as part of the study show that AI not only matches human intelligence but can even surpass it in certain tasks.
This achievement was made possible through a combination of learning and practical application. The system receives a word and is then prompted to apply it in a different context. For example, if the word “falar” (to speak) is given, the AI is asked to create contexts such as “falar muito” (to speak a lot), “falar pouco” (to speak little), “falar baixo” (to speak softly), and “falar alto” (to speak loudly).
As AI continues to evolve, it will be able to understand and utilize idiomatic expressions like “falar abobrinha” (to talk nonsense) and “falar besteira” (to talk nonsense). This expanded language understanding will enable AI to reach a much broader audience and communicate effectively with humans.
These advancements have promising implications for the field of programming. Computers will be better equipped to receive, understand, and act upon more complex commands. This opens up new possibilities for the development of sophisticated applications that can carry out intricate tasks with ease.
With the ongoing progress in AI and its ability to replicate human intelligence, questions arise about the potential future of machines. Will machines eventually reach a point where they possess all the capabilities and characteristics of human beings? While it is difficult to predict with certainty, the continuous advancements in AI suggest that machines could very well achieve complex cognitive abilities comparable to humans.
However, it is worth noting that human intelligence is not solely defined by our ability to make compositional generalizations. Our intelligence encompasses a wide range of skills, including emotional intelligence, creativity, and critical thinking, which are not yet fully understood or replicable by AI.
Although machines may continue to push the boundaries of their capabilities, it is crucial to acknowledge and appreciate the unique qualities that make us human. AI should be viewed as a tool that complements and enhances our abilities rather than replacing them.
In conclusion, AI has made significant progress in replicating human intelligence, surpassing certain aspects for the first time. With advancements like MLC, AI can successfully make composite generalizations, expanding its language understanding and potential applications. However, the full range of human intelligence is yet to be completely understood and replicated. AI should be seen as a valuable tool that empowers human capabilities rather than replacing them entirely.
