“Shall we play a game?” Who remembers the chilling computer voice that generated a question from the 1993 movie War Games which nearly started World War III? To prevent the computer from launching missiles, the cast had to teach it a lesson via Tic-Tac-Toe in which there were no winners. Our technology today has moved beyond that phase and is teaching computers how to communicate with each other – in their invented language.
Facebook recently had to shut down their systems because the newest artificial intelligence (AI) they were working on stopped speaking English and made up its own language – something humans can’t understand. The new program is an interactive system for users to be able to communicate as if talking to another human being. These chatbots are being designed to have the ability to plan conversations, not just answer simple questions.
During the experiments, the bots were paired with a human partner and they “reportedly managed to convince human participants they were negotiating with another person.” The bots are programmed to learn from their human contact and respond appropriately. However, during one of the experiments, two bots started communicating to each other in a language all their own:
In one exchange illustrated by the company, the two negotiating bots, named Bob and Alice, used their own language to complete their exchange. Bob started by saying “I can i i everything else,” to which Alice responded “balls have zero to me to me to me…” The rest of the conversation was formed from variations of these sentences.
In each exercise, the AI deviated from English to create a language of its own. Modern AIs operate on a reward system and expect a benefit when the task is completed successfully. Because there was no reward for using English, the AIs developed a more efficient form of communication. “Agents will drift off from understandable language and invent code-words for themselves,” Facebook AI researcher Dhruv Batra said. “Like if I say ‘the’ five times, you interpret that to mean I want five copies of this item. This isn’t so different from the way communities of humans create shorthand’s.”
It’s as if science fiction movies are jumping out and taking over society. The Matrix (1999) captured its audience by developing a world ran completely by computers. Humans were used as living batteries and the computers controlled everything perceived. In 2004 I, Robot pulled at viewers’ heart strings when one robot developed a consciousness and had dreams while the other robots went against their original programming and became assassins. Transcendence (2014) takes movie-goers even further after a brilliant authority on AI is killed, and his ‘brain’ is transferred into the mainframe of a very sophisticated computer system.
Now we have bots going against their programming and developing a new language that no human will be able to understand. Is this a good or bad thing? In some instances, it could be perceived as a good technological advancement. The ability for computers to communicate with each other means less human time spent programming and working with them. It could eliminate some programs that currently take years to develop.
But what happens if AIs start communicating with each other and leaving humans out of the equation? According to Digital Journal, “If AI-invented languages become widespread, they could pose a problem when developing and adopting neural networks. There’s not yet enough evidence to determine whether they present a threat that could enable machines to overrule their operators.” Batra also expressed his concern saying, “It’s important to remember, there aren’t bilingual speakers of AI and human languages.”
There are of course nay sayers out there who refuse to believe such a thing as computers taking over the human race could be even remotely possible. To a Doubting Thomas I say, remember the 1960s original Star Trek series? The cool gizmos and gadgets from the USS Enterprise seemed far-fetched and fantastical. Not many people imagined in just fifty years we’d have technology that nearly surpasses everything aboard that wondrous spaceship, and yet here we are. And never before has technologically advanced so quickly that new inventions and creations are barely even tested before they became old and replaced with something new and improved.
The repercussions of teaching AI to “think” for itself have not been fully researched, but humans are rapidly advancing into an anti-social society where everything is done via technology, including human interactions.
Will Facebook’s new technology lead to people forgoing even basic human nature while they turn to chatbots for friendship and solace? There’s something to ponder.
