Artificial intelligence (AI) is simultaneously fascinating and scary. We’ve all seen movies of rogue robots driving humanity to despair till someone finds a way to render them powerless. But we need to understand the term “AI takeover” before attempting to answer the above question.
What is an AI takeover?
The term refers to an imaginary situation in which artificial intelligence becomes a dominant force on Earth and establishes control over the human race. For example, AI may take over the work of all employees, resulting in a robot rebellion against humans.
Big names in science and technology, including Elon Musk and Stephen Hawking, have expressed their concerns about a possible AI takeover in the future. They have spoken about the importance of researching AI safety measures to make sure that any super-intelligent robots or computer programs that may be developed in the future remain fully under the control of their creators.
But are we really in danger? Can AI really take over our lives and create chaos on earth? Do we need to fear? These are few perplexing questions that are giving scientists sleepless nights at present. Nobody can deny the fact that AI is changing the way we live, and as the days pass, AI is getting deeper and deeper into our routine.
We cannot overlook the possibility that AI can harm as much as it can benefit humankind. For example, we cannot forget how Watson, a question-answering computer program developed by IBM, gave incorrect and dangerous treatment advice when asked to help medical practitioners with cancer patients.
Can humans create a super-intelligent robot?
To understand the answer to that question, we have to glance at the pages of history. Scientists coined the term “artificial intelligence” back in 1956 and used it to describe the simulation of human intelligence in machines or computers.
The world’s first electronic digital computers were developed just a decade before the term came into existence. Initially, the scientific community was excited and many believed that a super-intelligent robot would be created in just a few years, but nothing of the sort happened.
It was only in recent years that AI experienced a re-birth as scientists developed computer programs excelling in specific areas. For example, robots were created to defeat any human at chess, recognize the voice of a smartphone owner, find information online, drive cars, and so on.
This led to some experts predicting that a super-intelligent AI may be created with the power to perform multiple tasks. These robots will strongly resemble humans capable of utilizing available data for social manipulation. But such a super-intelligent robot has not yet been created.
Today’s artificial intelligence is weak or narrow as it is programmed to perform just one task, such as playing chess, driving a car, recognize faces, or collect information from the Internet. Computer programs are better than humans at performing the tasks assigned to them. But scientists and researchers have the long-term goal of creating a strong AI, which may outperform us in almost everything we do.
Can AI harm the human race?
AI may harm the human race if it is capable of experiencing emotions the way we do. But most scientists agree that a super-intelligent computer program cannot be capable of human emotions such as love, hate, anger, and so on. You cannot, therefore, expect the robot to be exhibit malevolence or benevolence with intention.
Here are two ways AI may harm humankind.
First, scientists can program AI to do something that may destroy entire nations or the entire world. For example, if autonomous weapons powered by artificial intelligence get into the wrong hands, they can be used to cause great harm. Also, a war involving artificial intelligence can result in widespread devastation.
Under certain harmful situations, it wouldn’t be possible to just turn off these weapons. As a result, human beings would lose control of their creation.
Second, scientists may program AI to perform a beneficial task, but the program may destroy life, property, and the environment while performing the task. For example, a car powered by artificial intelligence may obey all your commands, but may be found guilty of over speeding if you ask it to take you to a destination “as quickly as possible.” The “intelligent” car may even crash and kill innocent pedestrians in the process of getting you to the desired destination at the fastest possible.
Here is another disturbing example. Scientists may assign a geoengineering project to a super-intelligent system, but the system may cause harm to the environment in the process of completing the task. No human would be able to stop it because it would consider such attempts as threats to its purpose.
But are the above-mentioned possible? Many scientists think it is highly unlikely for the following reasons:
First, artificial super-intelligence or robots capable of behaving like humans isn’t going to happen overnight. It will happen very slowly, over several decades.
We have already started interacting with artificial intelligence and it has made our lives fun and easier. For example, we are familiar with Siri, Apple’s speech-recognition program; self-driving cars; and Watson, IBM’s question-answering robot, to mention just a few. At present, technology is capable of performing just a few simple tasks.
Also, the AI technology of today is smart and not intelligent as humans are. Robots and computer programs are incapable of reason, imagination, and adaptation. They cannot respond to changes in the environment, figure out the consequences of an action, interpret behaviour patterns, and think of possibilities the way we do.
It will take time and effort to create a machine capable of learning, imagining, and reasoning the way human beings do. It may happen in a few years, a few decades, or a few hundreds of years. Nobody knows when.
Right now, we have many predictions of human-made intelligent machines taking over the world and wreaking havoc on civilization. But the ones who make these predictions fail to understand that we have to first overcome the major research challenge of making a machine capable of human reason, imagination, and common sense before we can make such a machine for real. Super-intelligent AI will become possible only when we first solve this problem.
Conclusion
The biggest cause of concern at present is not the possibility of a future super-intelligent AI taking over the human race or rising against its creators. It is the way humans choose to use AI because the latter is just a tool that can be used in the right or wrong way. We have to fear not the intelligence of robots but the unintelligence of human beings.