Facebook’s scientists have been forced to shut down an artificial intelligence program after realizing that it has created its own distinctive language that humans cannot comprehend. This follows after scientists at the research lab discovered that the chat-bots developed their own code words to make communication more effective without human input thus deviating from the script. It is as worrying as it is incredible – simultaneously a glimpse of both the amazing and surprising capability of artificial intelligence. However, what we need to understand is that artificial intelligence and machine learning possess the phenomenal ability to accelerate, simplify and advance many aspects of our lives.
Computers are capable of processing and ingesting huge amounts of data and extracting patterns and important information at a rate exponentially faster than human and it’s that capacity that is being explored all over the world. This comes after a number of scientists and tech luminaries including Stephen Hawking, Elon Musk, Bill Gates have cautioned that artificial intelligence could cause disastrous unexpected consequences. However, Facebook CEO Mark Zuckerberg condemned the warnings terming them as “petty irresponsible” warnings from pessimists.
Renowned physicist Steve Hawking warned in 2014 that artificial intelligence could mean the end of human race. “It would take off on its own and re-design itself at an ever increasing rate. Humans, who are limited by slow biological evolution, couldn’t compete, and would be superseded.” In addition to that, billionaire tech entrepreneur Elon Musk warned that artificial intelligence risks an “existential threat” to human civilization. He said that the advancement of the self-governing weaponry could lead to an “AI arms race” therefore it’s impacts could be so great and hence needs to be proactively regulated before it’s too late.
Whether he insinuated it or not, in some people’s minds Musk’s caution invoked images of Skynet, a fictional Artificial intelligence network in the Terminator film series that developed self-governing techniques and set out to terminate the human race. However, this is just one of the many documented cases of AI developer’s encounter with the mysterious artificial intelligence behavior and in all cases, the artificial intelligence system deviated from its training in English to develop a new language that cannot be understood by humans.
The phrases in the AI developed language makes no sense to humans, but contain vital information when interpreted by AI bots. For example, at Elon Musk’s open AI lab, an experiment succeeded in having AI robots build their own code words. Also, at Google, a group of researchers working on the translate service revealed that the artificial intelligence they programmed had silently written its own language to aid in translating sentences. The translate developers had added a neural network to the system, making it capable of translating between language pairs it had never been taught.
The new language the AI silently wrote was a shocker. This may sound like a groundbreaking innovation, but in reality, it offers very little real-world value. The human brain is incapable of comprehending the conversation. Despite the fact that it is more effective and faster way for AI robots to communicate with each other, the language is of little importance to humans who are working tirelessly to optimize AI-to-human speech not AI-to-AI. The world is wired and connected.
The AI will finally figure out how to cooperate with other AI systems and maybe they will realize that humanity is a threat or an inefficient waste of resources- assumptions that seem reasonable from a purely logical view. It is not the end of the world. And am not insinuating that we need to do away with machine learning and artificial intelligence and embark on other innovations. All we need to do is to proceed with caution.
We need to closely monitor and understand the self-perpetuating evolution of an AI system, and always have some means of controlling it. This is because if we don’t understand its language, we may not even be able to establish how or why it behaves the way it does, and may not work out well for mankind.
Furthermore, there is no solid proof that these unexpected AI deviations are a threat or that they could lead to machines taking over developers. They do make advancements challenging, however, because people cannot understand the amazingly logical nature of the new languages. Facebook hopes it can optimize its artificial intelligence into a personalized digital assistant for future services.
Should we be worried about the future of artificial intelligence?
Is it really a beneficial invention to humans?