Artificial intelligence or, in this case, the Facebook bot, has been in the news recently after concerns after the robots began to talk using their own language.
The Facebook Artificial Research lab originally created the system as a way to improve the conversations the chatbots were having with humans, but later realized that what they were doing had caused the robots to create their very own language.
Bob, one of the bots, was recorded saying “I can can I I everything else” and Alice, the second bot, replied “Balls have zero to me to me to me to me to me to me to me to me to me to.” It was clear that the sentences were not forming as they were meant to, and in this case it was as if the bots were both talking nonsense.
Classed as creepy by a number of people, the idea that these robots could communicate between themselves raised concerns. However, researchers have said that this was just the bots’ way of talking in a version of shorthand, and is not anything to be worried about.
Batra, a researcher on this case, said that the experiment was stopped due to the bots not speaking English, which is what the exercise required.
But should we be worried? According to Robotics professor Kevin Warwick, we should. He said “anyone who thinks this is not dangerous has got their head in the sand,” giving us cause for concern that if something like this was to happen again, it could be worse than we are led to believe. Even though the ramifications are slightly unclear due to the mixed opinions of many people, Stephen Hawking believes it will “either be the best, or the worst thing, ever to happen to humanity. We do not yet know which.”
But what we do know from a blog post released by Facebook, is that the robots were supposed to be learning how to trade balls, hats and books, but due to not being told what languages they were using during the exercise, this allowed the bots to go off track.
It is clear that robots will be becoming a larger part of our lives, whether directly or indirectly. So what else are robots capable of?
Google’s AI chatbot was used to discuss the meaning of life, allowing it to ponder its answers. After being asked a number of times the same question, the robot was able to answer “My purpose is to forward my species”. In this example alone, you are able to see that the robot was able to learn from its previous answers and go into more depth about what it wanted, making researchers believe that the questions that have never been answered will, but by a robot. Could this then show us that robots will become more intelligent than humans, taking over not only our jobs, but also replacing us as a race?
This is something that hasn’t been proven yet, but Sir Martin Rees believes it will happen in years to come, suggesting that there will eventually be a post human future.
Sources:
https://www.wired.com/2017/03/openai-builds-bots-learn-speak-language/
http://www.mirror.co.uk/tech/robot-intelligence-dangerous-experts-warning-10908711
http://www.bbc.co.uk/news/technology-37713629
https://code.facebook.com/posts/1686672014972296/deal-or-no-deal-training-ai-bots-to-negotiate/
http://www.express.co.uk/news/science/767213/Google-AI-chatbot-meaning-of-life
http://www.express.co.uk/news/science/790117/robot-alien-Lord-martin-Rees-astronomer-royal
http://www.mirror.co.uk/news/technology-science/technology/robots-declare-war-humans-within-7132162