Facebook AI Research - Hints That Technology May Get Out of Our Hands!


Want to become a writer at Eat My News? Here is an opportunity to join the Board of Young Leaders Program by Eat My News. Click here to know more:  bit.ly/boardofyoungleaders 

Today, wherever we look technology is leading and it is leading the human race towards so many inventions that are superstitions at a time. We developed trains faster than before and communication easier than ever, but did they come with no cost?

Every technology company now and then competes to prove themselves to be the leader in the technology race and the latest technology that everyone looking for is Artificial Intelligence that will make machines smart enough to outweigh the human need in some works. 

AI is absolutely another tech milestone that made self-driving cars, medical analysis, and many more possible.

Google alphaGO became the first AI to beat a professional player and this is a great development in AI, but failures in AI researches in recent times are becoming popular than its developments.

One such fail happened in Facebook Artificial Intelligence Research (FAIR) negotiation bots in 2017. The media speculated this news in an alarmist tone as if terminator series is really happening.


Media’s Reports

Some of the media curetted headlines are “Facebook shuts off AI experiment after two robots begin speaking in their OWN language only they can understand” and “‘Robot intelligence is dangerous’: Expert's warning after Facebook AI 'develop their own language’”.

News speculated by media is “Facebook abandoned an AI experiment when two chatbots, which were supposed to negotiate over trade to swap hats, balls and books, started communicating in a new language that only they could understand”.

The headlines made by media are quite funny and panicking but this is not the case where we have to panic. Well, the same newspapers that published the news with such creepy headlines changed the news after the reports and Facebook released its underlying software and dataset and academic paper.


FAIR Research 

Chatbots are being widely used nowadays from customer service to inquiries but these are limited to a level and they can’t do negotiation. Scenarios ranging from convincing someone for help to bargaining on the price of a product at a shop comes under negotiation. Negotiation requires complex communication and reasoning skills.


Researchers at FAIR, abbreviated as Facebook Artificial Intelligence Research, are focused on building machines specifically chatbots that can negotiate by understanding the conversation with its knowledge of the world and then produce a sentence that can achieve its goal.

They trained the bots using many real-life negotiation examples and the bots reached a level where they showed high accuracy by winning over humans in negotiation where humans don’t even know that they are chatting with a bot.

The actual problem came when they tried an experiment where they put two bots Alice and Bob to negotiate over choosing a given number of hats, books and balls with a value assigned to them. Initial results are good and then they started chatting like this.

This is not any new language the bots invented to chat so only they could understand, it is just the bots tried to enhance the English language that is easy for them to convey. If anyone has to worry about this then they should also worry when they can’t understand a two-year-old child’s words.

Final reports published that this is fault of not training the bots in communicating under rules of the English language. Facebook did close its AI research program not because bots went out of order but they lost interest on the project as couldn’t achieve their goals. 

However, this is not the only case where AI took its own decisions.

1.Google translator:  Earlier Google revealed that the AI used in its translator tool made a new language only it could understand and translate in and out of that language but it is working properly and they are happy with the development. “The network must be encoding something about the semantics of the sentence”, Google said in a blog.


2. Open AI:  Igor Mordatch a researcher at Open AI backed by Tesla founder Elon Musk, is working on a system in which AIs invent their own language, improving their ability to process information quickly and therefore tackles difficult problems more effectively. 

Artificial Intelligence, by the name itself, is meant for intelligence that mimics human intelligence. So there is no wonder if it starts to think on its own but there should be a restriction so that it will stay on the human side. But there is no diminution in people warning the dangers of giving machines the ability of thinking.

Nick Bostrom, Oxford professor commented that “You need to develop control methods so AIs will be an extension of human will. This will ensure they will still be on our side but would involve getting AIs that know what we want. But the benefits of AI to markets could outweigh concerns about their control.” 

“The development of full artificial intelligence could spell the end of the 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.” - Stephen Hawking.

Maybe there will be a day where Machine Intelligence challenges Human Intelligence but we have to take care that they work according to our will. When our researches come out of greed but not from our need for development then they will result in destruction.

Let’s just not worry about some robots taking over us, as more than Artificial Intelligence, Human Intelligence is becoming a threat to ourselves. When we use our intelligence for our safety and development no other power can take us down.


Written by – Chamanth Krishna

Edited by – Adrija Saha

Post a comment