In 1984, Hollywood introduced us to the US military’s global artificial intelligence (AI) project, Skynet – a system that, in the interest of self-preservation, led to humans being exterminated in order to create an overpowering, global artificial intelligence hierarchy. The system that was originally activated by the military to control the national arsenal soon begins to learn at an exponential rate and gains artificial consciousness, which seeks to destroy humanity.
The Terminator series is purely fictional, of course. Or maybe not! With accelerating technological development, one can’t help but envision a future where machine intelligence will take over from human intelligence. Today’s artificial intelligence powered systems could evolve into super intelligent machines that will be self-aware, create their own agenda, and might even be able to create far more intelligent copies of themselves. That is when technological singularity will occur.
Just how smart will AI be allowed to get until they are stopped?
In July 2017, Facebook carried out an artificial intelligence research project involving two chatbots, which required them to communicate with each other in English. The interaction began in the conventional manner, where the bots first established a link before communicating. What had started as a normal exchange, soon turned into something completely unexpected. The robots began to communicate in what appeared to be nonsensical messages. After a closer look, however, researchers noticed a peculiar pattern hidden in the communication, which turned out to be a carefully structured secret language that only the machines could understand.
Facebook immediately shut down its system and terminated the experiment, as the chatbots had deviated from the laid down course and were communicating in a new language developed without human aid.
Earlier in 2017, Google revealed something very similar to Facebook’s experiment. The AI that the company uses for its Translate tool had created its own common language, which is not readable by humans. However, unlike Facebook, Google seemed happy with the development and allowed it to continue.
Should we be worried?
Regardless of the probable implications, the development and adoption of artificial intelligence is accelerating in most parts of the world. Companies are investing billions of dollars into the technology, and their phenomenal impact on our lives is unlikely to be halted anytime soon.
While AI is not all-pervasive yet, it has undoubtedly become part of our daily lives. As the Internet of Things (IoT) continues to grow, our interactions with artificial intelligence will only increase. However, it might be to too early to start worrying about their hypothesized existential risk to humanity.
With most AI systems having only virtual existence, the manifestation of a self-replicating machine seems implausible. Secondly, human-like AI, which is essential to achieving singularity, is still a distant future. As intelligent systems are essentially based on models of the neocortex only – the part of the human brain that has the ability to learn languages, computer programming, and design buildings and more – they lack the instinctive behavior and emotions that define the very essence of humanity. Intelligence is not something that can be controlled by the press of a button. It requires continuous manipulation in addition to an advanced brain.
Machine intelligence, for now, does not represent a threat to humanity any more than computers do. What it does do is provide the opportunity for humans to extend their time on this planet by generating new knowledge and working together to make our lives better.