It’s no doubt that Artificial Intelligence (AI) is advancing at a phenomenal rate. Just a few years back, there was virtually no talk of mainstream AI applications and now it seems that everyone is talking about it.

Such a rapid progress means that general purpose AI, that we have been seeing in sci-fi movies like Iron Man, could be here sooner than we think. Whether or not that is a good thing, is a whole different story.

Right now, the research on AI is focused on giving it the capability to think and learn like humans. It’s like, building the future AI in our own image, only much more intelligent and capable.

This is exactly what an AI system, created by researchers from OpenAI achieved. The surprising part is that they didn’t even teach the AI how to do it. It learned it by itself!

The Sentiment Reading AI

OpenAI is a non-profit AI research company which boasts investors like Elon Musk, whose mission is to develop a friendly AI so that humanity doesn’t meet the same fate as is so often showcased in the movies. Elon Musk had once stated that he had joined another AI research company “DeepMind” to “keep an eye” on the development of AI.

Anyways, researchers at OpenAI were experimenting with an artificial neural network whose goal was to predict the next character in the text of reviews that people leave on Amazon. For this purpose, they had trained the network and fed it tons of data over a period of a month.

To the surprise of the researchers however, the AI system also learned how to analyse sentiment by classifying the reviews as positive or negative with an accuracy of 91.8 % which is pretty accurate for an artificial system!

So, basically while performing a task of predicting the text in reviews, the AI learned to differentiate between positive and negative reviews based on the words that were used in the reviews. Now that is amazing!

The Future

In the future, AI will be so advanced that it will basically learn just like humans. We learn new things all the time and we choose what we want to learn and how we want to learn it.

In the future, AI will be able to do the same. Many experts have warned against this and rightly so. What if an AI system learns to kill humans or other organism in the world and “liked it”?

It could quickly spell the end of humanity and possibly all life forms on Earth and it wouldn’t be something we just consider to be under the purview of Hollywood but a real life threat.

That’s why it’s of vital importance that everyone involved in the development of futuristic AI system do so, keeping in mind that humanity’s existence may very well be in their hands!