Even though human level AI is still a step away just yet, researchers in artificial intelligence are looking into new ways of teaching machines to know their limits. This is the case with the research conducted at Spain's Artificial Intelligence Research Institute.
The film 2001: A Space Odyssey (1968) was released nearly half a century ago and Transcendence has hit the silver screens just last year (2014). They both are presenting related different dramatizing variants of the same story about a scientist's consciousness being uploaded into a computer.
Their plots, despite being half a century apart, are broadly similar. The arrival of human-like machines that rebel against their creators continues to be a preferred theme of science fiction stories.
However, coming back to reality, we can notice that over the last 30 years, progress has been slower than expected in the field of artificial intelligence research. While artificial intelligence has become a part of our everyday lives, in our cars and phones, and today's computers are able to process large amounts of data, AI systems still lack human-level ability to use the information provided in order to make deductions.
While humans can read and understand different sections of a newspaper, grasping the implications and consequences of a story, machines are still not able of this performance. Humans can acquire experience just by interacting with their environment, achieving tacit knowledge. Today's artificial intelligence systems don't have that kind of ability yet.
Common sense reasoning is still to be achieved by AI machines. This is one of the challenges in artificial intelligence research. According to the director of the Spanish National Research Council's Artificial Intelligence Research Institute, Ramon López de Mántaras, our machines can perform very well at complex tasks such as playing chess, but they are specialized and cannot play domino too, for instance.
The director of the Spanish AI Research Institute participated in a recent debate organized by the Catalan ministry of ITC. According to Ramon López de Mántaras, AI research has focused in the last 30 years on weak artificial intelligence, making machines very good only at a specific specialized topic or task. However, artificial intelligence systems have not progressed that much in being able to achieve common sense reasoning.
This current phase in development of AI systems may change soon with the development of cognitive computing systems, smart cities and smart buildings.
The AI systems designed for these applications will have to carry out human-like analysis of complex and diverse data sets. However, López de Mántaras, tends to disagree. He expressed his opinion that high performance or big data are not the factors bringing us closer to advances in AI.
Unlike the skeptical López de Mántaras, those futurists who talk about "the singularity", believe that the advent of artificial general intelligence known also as strong AI, will occur not later than the period between 2030 and 2045.
According to their predictions a big advancement in computer science is going to happen in the near future. Non silicon-based technologies such as the new memristors, or DNA computing will facilitate this big step forward.
