Teleporting imagination to reality

Be it RoboCop in the 1980s or The Hitchhiker’s Guide To The Galaxy and I, Robot in the 2000s, artificial intelligence has been an efficient and popular tool for good science fiction. It wasn’t long ago that we called on-screen AI “futuristic". Before we realized it, AI had touched our lives, and in ways we could not have imagined. 2016 was the year it really hit closer home that AI is here to stay.

Have you ever used AI?

Yes, you have. Deep neural networks have changed the game rapidly. Facebook and Google can recognize the faces of your friends in the photographs you post, and understand the context behind the commands you shout into your smartphone. Those website and search suggestions that you see as you type in the address bar in your Web browser, how Microsoft Outlook mail filters some new messages to the Clutter inbox, Amazon’s algorithm, which suggests the books you should read, the way Apple’s Siri understands your requests, and the way your phone provides suggestions for the next word as you type a message—these are all examples of AI at a very basic level. They are becoming irreplaceable fixtures in personal technology as we try to be more productive.

This year, AI truly evolved. Personal assistants for your smartphone are a form of AI that has an impact on most users right away. A journey that started with Apple’s Siri assistant many years ago has now added Google’s Assistant, Microsoft’s Cortana and even Amazon’s Alexa, into the fold of intelligence. And there is more. “We believe that AI has the potential for transformative, positive impact in the world. Fulfilling this potential is not only dependent on the quality of the algorithms being engineered and the data they use, but on the level of public engagement, transparency, and ethical discussion that takes place around them," said Mustafa Suleyman, co-founder and head of applied AI, DeepMind, Alphabet Inc.’s artificial intelligence division, earlier this year in an official statement.

Where did it all begin?

As early as the 1950s, Alan Turing, a British mathematician, was grappling with a conundrum: Can machines think? This was a simpler time, and the term “artificial intelligence" was still a million miles away from being coined. Later, it became known as the Turing Test, which meant replacing a human with a computer program. The test involves a man and a woman who are being questioned by an interrogator. The interrogator has to guess the gender by asking questions and analysing the written replies. In this, the woman is trying to help the interrogator, while the man is trying to complicate things. Turing replaced the human interrogator with a computer. He summarized his objective: “What will happen when a machine takes the part of A in this game? Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, can machines think?" Experiments, such as those by American cognitive scientist Marvin Minsky, carried on but it was many years later that the true power of AI was recognized.

How is AI shaping what we do?

AI works on one very simple premise—the ability to learn continuously and improve its algorithms, and try to include contextual information in responses and suggestions. Depending on the particular AI’s implementation, it learns over time, and tweaks its behaviour. That is the thing with AI—there is no one size that fits all. For example, your Netflix suggestions will be based on what you’ve been watching over the past few months, and these will be different from what I see. When Google’s Assistant is invoked in a text chat in the Allo app, it tries to understand the content of the conversation from both parties, and not just one, to generate suggestions.

AI as a licensed driver

In the year 2015, Denis Sverdlov, chief executive officer, Kinetik, announced a joint venture with Formula E to create the Roborace self-driving championship. The idea was to feature 20 identical cars allocated to 10 teams; the series would run on the same circuits as Formula E. The unique aspect—there would be no racing drivers, and the cars wouldn’t be remote-controlled either. The fully autonomous racers are based on the NVIDIA Drive PX 2 supercomputer to run the algorithms. And it’s not just racers.

The cars that you may drive in the future may be based on machine-learning algorithms. After the well-documented crash of a Tesla car, which resulted in the first-ever autonomous car accident fatality, Elon Musk, CEO, Tesla, said, “Full autonomy is going to come a hell of a lot faster than anyone thinks it will..." and went on to talk about the “narrow AI", which makes crucial decisions and is capable of planning routes without any driver interference.

Businesses want to be ready for the opportunities that may arise with new intelligence. For instance, ride-sharing service Uber has completed the acquisition of AI start-up Geometric Intelligence, founded by New York University psychologist Gary Marcus and University of Cambridge professor of information engineering Zoubin Ghahramani. Uber already has a self-driving car lab in Pittsburgh, gets assistance from carmaker Volvo, and has also recently acquired self-driving car company Otto. It doesn’t need a rocket scientist to put the pieces of this puzzle together.

The small matter of doomsday

With each advance in software and algorithms, we are only too happy to hand over a part of our abilities to machines. The illusion is that these machines are doing tasks as well as humans, if not better. This is freeing up humanity to focus on other perhaps more desired tasks, though that remains debatable. Certainly, the progression towards an algorithm-driven humanity is moving along rapidly.

In some ways, while we are hurtling towards the unknown, the risks seem to have paid off for the moment. No discussion about AI can be complete without a slight left turn towards visualizing potential doomsday scenarios, spearheaded by machines that have a mind of their own, perhaps the kind that the Terminator movies had envisioned. Solace lies in that AI still needs human inputs, such as chatbots and concierge apps.