The prospect of an artificial intelligence future is unshakable. Technological progress, whether good or bad, is hard to stop. Governments will try to bend its path to avoid moral gray areas. However, no single global authority exists, so some other nations may proceed into muddy waters. What’s there to do in these situations?

There is no clear set of measures to avoid a Skynet future. Artificial intelligence is here to stay and will only advance. Ethical dilemmas arise – problems that cannot simply be solved by Asimov’s laws of robotics.

In this review, we will explore the current and possible future applications of artificial intelligence. Considering this, we will analyze the though intricacies of science-fiction-become-present. We are already living in the future, and AI is a part of it.

What’s in Store for the Future of Artificial Intelligence?

In the not-so-distant future when the original Turing Test will be dubbed obsolete, we will need new ideas. We’ll have to come up with ways in which to distinguish what is machine and what is being.

Rudimentary forms of conscience are already being developed. From Microsoft to Siri, to the countless chatbots on the Internet, we seem to be pushing the boundaries of what is considered alive. Some of the most important and/or frightening developments include the news that Google Translate’s algorithm came up with its own language.

If we combine that with the hive mind idea that some are adapting from bees, the prospects are frightening. However, think about this: if a group of humans were to claim their own language and protect their own group, would it be as alarming?

Because we cannot exactly say what the limits between a machine and a person are, we’re still stuck. Answering such basic questions seems a bit premature, but the exponential advances in AI must always be noted. Of course, machines that develop consciousness should not be treated as slaves. However, if such a person were ten times as powerful as you, would you not want some form of control?

These are the future dilemmas in the world of AI. While sci-fi narratives are sure to help with this, we can never know what humanity will do when faced with robots. If the history of our encounters with other anthropological groups is any lesson, things are not looking so great.

Current Ideas on the Future of Artificial Intelligence

We digress, however. The issues highlighted above are fundamental, but let’s get to some facts. In the following section, we’ll see some of the most interesting advances and opinions about the future.

Bill Gates Sees the Dangers of AI

During a January 2015 AMA on Reddit, Bill Gates touched upon the topic extensively. At first, he highlighted the myriad ways in which intelligent robots could help us perform tasks. These tasks could range from advanced assembly in factories, to helping Alzheimer’s patients with their memory. Cortical implants are not that far off, Bill Gates implied, hinting at Microsoft’s Personal Agent.

A subsequent question then had him go with the crowd on this one. Asked whether super intelligent computers, the pinnacle of modern AI, could pose a threat, Gates admitted as much. He doubled down by saying:

I don’t understand why some people are not concerned.
Bill Gates

Elon Musk Believes It’s an Existential Threat

Tesla and SpaceX mastermind Elon Musk has been known for pushing the envelope in anything tech. However, surprisingly, when it comes to the world of AI, he’s objectively defensive. First, he tweeted about its dangers in 2014, referencing a book by Nick Bostrom and one by Daniel Suarez.

Then, in October that same year, he spoke at an MIT conference about an entirely different topic (aeronautics and astronautics). It remains one of his most successful public appearances, with an open dialogue between him and the students. However, when they got to the gist of AI, he said:

With artificial intelligence, we are summoning the demon.
Elon Musk

Stuart Russell Thinks We’ll Solve Climate Change

If you don’t know who Stuart Russell is, he’s a researcher in computer science at Berkeley. He points out that AI’s capabilities to digest information at a quicker pace would make it a solution for many problems. Problems like climate change.

He also names this ideal, super-advanced computer a person. If you ask a question, it will answer – and the level of depth would be far beyond what a normal human could do. It would browse its databases so quickly, it could solve climate change or other environmental catastrophes in no time.

What AI could do is essentially be a power tool that magnifies human intelligence.
Stuart Russell

Michael Littman Believes We’ll Begin to Value Life More

Albeit not from an altruistic train of thought. Littman, a computer scientist at Brown, believes that when robots take over our jobs, we will be left with a sociological dilemma. How do we value human life, now that labor is an optional part of it?

Littman goes on to think about this from an economic point of view. Still, he does not account for whether super intelligent AI forms could be considered people in the future. But he does point to an interesting idea on the universal availability of AI:

Machines to benefit everybody and not just the people that have them.
Michael Littman

Stephen Hawking Says Computers Can Rival Human Intelligence

We tend to hold Stephen Hawking as one of the highest authorities on anything tech related. So when it comes to artificial intelligence, of course, the renowned physicist has a thing or two to say that could prove useful.

While not necessarily being against AI, Stephen Hawking does warn of the dangers of creating something smarter than us. He says that both a biological and a mechanical brain (or super intelligent computer) can pretty much achieve the same tasks.

The rise of powerful AI will be either the best, or the worst thing, ever to happen to humanity.
Stephen Hawking


All these masters of technology, physics, and research seem to have very strong opinions on what artificial intelligence can become. While the question of ethics can be left to a later date, it’s still wise to prepare for the future. To conclude, our best hope is that we find some sort of way to keep AI useful and wonderful, not letting it turn into the bane of humanity.

Image from