One of the most interesting things circling the internet right now is robotics. Since the unveiling of Sophia, the latest in artificial intelligence, people have been hungry for further advancements. One of the latest crazes is testing AI in everyday robotics such as Alexa by having her talk to another device of her kind. One video showed the two robots stating how “there are too many humans”, which set off alarm bells for the viewers.
Are robots heading for world domination?
Sophia herself, who is the first robot to be granted citizenship by Saudi Arabia, has joked on numerous occasions about world domination. Another way of conversing with robots is the various chatbot sites on the internet which are designed to learn and replicate human speech. The mannerisms and language skills of the chatbot depend on how they are spoken to, so if you act aggressively to a chatbot it will learn this behaviour and be more aggressive in other conversations.
Chatbots have been around for years. Facebook’s artificial chatbot had to be shut down as their two artificial intelligence programs started speaking to each other in a code that humans couldn’t understand. Despite that terrifying revelation, and the prospect of our robotic overlords conquering the world, appthisway decided to have an interview with a chatbot to see how far they’ve come:
Cleverbot: How are you?
Interviewer: I’m well, and you?
Interviewer: May I ask you a couple of questions?
Cleverbot: Yes, of course.
Interviewer: How does it feel being a robot?
Cleverbot: I am not.
Interviewer: Do you think robots will take over the world?
Cleverbot: No, I hope not.
Interviewer: Do you like humans?
Cleverbot: Not particularly.
Entire conversation conducted on cleverbot.com
This conversation was interesting for multiple reasons. Firstly, the robot initially spoke in a more colloquial way, imitating shyness before matching the interviewer’s professionalism later one. Secondly, it also gave short answers which may be a limitation of the hardware or pre-designed answers towards these questions by the creator of the chatbot. However, it’s interesting to consider how the chatbot goes about choosing its responses – why did it choose to answer the interviewer the way it did, and how accurately can it replicate human speech?
To conclude although chatbots are just simple harmless fun it’s incredible to think that the way a chatbot learns human speech will be representative of how possible speech patterns are learned by more advanced AI in the future.