Today, many global technology companies are participating in a unique race: they are striving to literally breathe life into artificial intelligence (AI). Machine learning systems have already become an integral part of the business, so it's no wonder that you can come across the news about AI and neural networks almost every day. Are we really much closer to the moment when the intelligence of machines can be compared to the human, or to the moment when a person and a machine can conduct a small talk and work together as naturally as people do among themselves? Are the machines far from gaining self-consciousness?
People like Yann LeCun, the director of the AI development department in Facebook and the professor of computer science at the University of New York, believe that we overestimate the capabilities of today's AI systems.
"In fact, we are still far from creating machines capable of learning the basic perception of the world in the same way that humans and animals can do it," LeCun commented in an interview with the portal The Verge.
"Yes, you cannot argue that in some areas machines already acquired capabilities superior to human, but in the perspective of creating a common universal artificial intelligence, we did not even approach the level of the rat."
The so-called common artificial intelligence is a system that does not require the participation of a human operator and is capable of performing practically any task that a person can perform. Current AI systems are specialized and can work only with one task, for example, to engage in speech or images recognition or to highlight specific objects in huge data arrays, and that is just what they were programmed to do.
Manuel Cebrian, a staff member at the Massachusetts Institute of Technology and one of the developers of Shelley, an AI algorithm, capable of writing terrible stories agrees LeCun’s statements.
"AI is just a great tool. But in my opinion, based on my experience with Shelley, the AI is far from the ability to write peace in the genre of horror at the professional level, since it is still far from the level of human intelligence," Cebrian says.
LeCun generally believes that, despite all the amazing levels of progress at the AI researches that researchers have been able to reach over the past few years, it wasn’t exactly the development of the very true artificial intelligence that everyone dreams of today.
"I do not in any way want to humiliate the merits of our DeepMind engineers and researchers working with the same AlphaGo, but when people interpret AlphaGo's improvement as a significant advance in the development of the common artificial intelligence, it is wrong. Because it's not at all like that," LeCun commented.
Piero Baro, CEO of Aiva Technologies, which developed the AI algorithm for creating music, also believes that the successe we have achieved in creating synthetic intelligence is somewhat exaggerated.
"The Common artificial intelligence is a topic that attracts a lot of attention. I'm optimistic about rapid technology development, but at the same time I think that most people simply do not understand the complexity of our own brain, not to mention how difficult it will be to create an artificial one," Baro said.
Today, people like to use the term AI for any reason. Terms such as "machine learning" or "deep learning", as well as "neural networks" can be found in any AI news. Although each of these terms is somehow associated with AI, in fact, the speech here is not about the AI as such.
Machine learning is an instrument. A set of algorithms that make up the intelligent system, which is trained by absorbing a huge amount of data. Also, in-depth training is a type of machine learning that is not necessarily tied to a specific task. On the other hand, the neural network is a system that simulates the work of the brain, but again it works only within the framework of those features that create algorithms for machine learning.
AI experts believe that all three of the above-mentioned components are the fundamental basis for the creation of synthetic intelligence with the ability to think humanly, that is, to realize their actions and their consequences. But we are only at the very beginning of this path.
We have really made great progress, but the current developments practically haven’t shifted towards the creation of real intelligence. Nevertheless, it is quite interesting to ask the question of when we should expect the appearance of this type of artificial intelligence. Are there any time frames?
According to Luc Tang, the head of the TechCode AI startup, a real shift towards the creation of a full-fledged artificial intelligence "will begin with a breakthrough in the development of uncontrolled machine learning algorithms." As soon as we come to this, "machine intelligence will quickly surpass humanity," Tang shared in an interview with Futurism.
"To create a full-fledged common artificial intelligence we need significant progress in the development of neuroscience and the development of hardware, " Baro says.
"We are standing at the edge of Moore's law. Transistors become so small that it's simply impossible to get them physically. And new hardware environments such as quantum computing, have not yet been able to demonstrate their superiority over our conventional hardware, even when performing standard tasks, "adds the expert.
Many agree that to consider AI as the real intelligence, it must cope with five tasks, the first of which is the Turing test, where the machine needs to convince the person that the person is talking to another person, not a machine. Baro is convinced that the present generation will witness AI successfully passing the Turing test. Nevertheless, the expert believes that "this will not necessarily be the common artificial intelligence, but something that is already close to it."
It should be mentioned that common artificial intelligence will be a precursor to the so-called technological singularity - the moment when intelligent machines will surpass the human level of intelligence, stimulating the rampant and exponential technological growth, which promises to transform our lives at the fundamental level. In 1993 Vernor Vinge, the author of the term and the concept wrote the following:
"Soon we will be able to create intelligence that will surpass our own. When this happens, human history will reach a kind of singularity, an intellectual transition to a new level. From that point, the world will begin to change so much that it goes beyond our understanding. "
However, Ellon Musk, Stephen Hawking and even Bill Gates aren’t very pleased with this prospect. They state that people simply do not understand what actually means acquiring artificial super-intelligence, and we obviously are not prepared for those possible consequences that technological singularity can bring.
Why is it necessary to consider artificial intelligence as a fall for all mankind, and not as its companion, friend, helper, after all? Musk, for the sake of justice, considers this idea, so he created the project Neuralink. Kurzweil mentioned this kind of cooperation between man and machine when he said that in the future nanobots will inhabit us, which will greatly improve our capabilities.
Algorithms like Aiva and Shelley already have shown their usefulness in working with people. At the same time, intelligent robots like Sophia from Hanson Robotics and Pepper from SoftBank easily allow us to at least imagine that truly intelligent machines are already among us.
"We will reach the technological singularity, perhaps in 30-50 years. On the one hand, it may seem that this is very long, but on the other - it means that many of us will have a chance to live up to this point, " Tang summed up.