Investing in AI offers more rewards than risks

By KR Sanjiv for TechCrunch

It’s difficult to predict how artificial intelligence technology will change over the next 10 to 20 years, but there are plenty of gains to be made. By 2018, robots will supervise more than 3 million human workers; by 2020, smart machines will be a top investment priority for more than 30 percent of CIOs.

Everything from journalism to customer service is already being replaced by AI that’s increasingly able to replicate the experience and ability of humans. What was once seen as the future of technology is already here, and the only question left is how it will be implemented in the mass market.

Over time, the insights gleaned from the industries currently taking advantage of AI — and improving the technology along the way — will make it ever more robust and useful within a growing range of applications. Organizations that can afford to invest heavily in AI are now creating the momentum for even more to follow suit; those that can’t will find their niches in AI at risk of being left behind.

Risk versus reward

While some may argue it’s impossible to predict whether the risks of AI applications to business are greater than the rewards (or vice versa), analysts predict that by 2020, 5 percent of all economic transactions will be handled by autonomous software agents.

The future of AI depends on companies willing to take the plunge and invest, no matter the challenge, to research the technology and fund its continued development. Some are even doing it by accident, like the company that paid a programmer more than half a million dollars over six years, only to learn he automated his own job.

Many of the AI advancements are coming from the military. The U.S. government alone has requested $4.6 billion in drone funding for next year, as automated drones are set to replace the current manned drones used in the field. AI drones simply need to be given a destination and they’ll be able to dodge air defenses and reach the destinations on their own, while any lethal decisions are still made by human eyes.

“The prevailing wisdom is that the risk of being left behind is far greater than the benefits of playing it safe.”

On the academic side, institutions like the Massachusetts Institute of Technology and the University of Oxford are hard at work mapping the human brain and attempting to emulate it. This provides two different pathways — creating an AI that replicates the complexities of the human brain and emulating an actual human brain, which comes with a slew of ethical questions and concerns. For example, what rights does an AI have? And what happens if the server storing your emulated loved one is shut down?

While these questions remain unanswered, eventually, the proven benefits of AI systems for all industries will spur major players from all sectors of the economy to engage with it. It should be obvious to anyone that, just as current information technology is now indispensable to practically every industry in existence, artificial intelligence will be, as well.

The future of computation

Until now, AI has mostly been about crafting preprogramming tools for specific functions. These have been markedly rigid. These kinds of AI-based computing strategies have become commonplace. The future of AI will be dependent on true learning. In other words,AI will no longer have to rely on being given direct commands to understand what it’s being told to do.

Currently, we use GPS systems that depend on automated perception and learning, mobile devices that can interpret speech and search engines that are learning to interpret our intentions. Programming, specifically, is what makes developments like Google’s DeepMind and IBM’s Watson the next step in AI.

DeepMind wasn’t programmed with knowledge — there are no handcrafted programs or specific modules for given tasks. DeepMind is designed to learn automatically. The system is specifically crafted for generality so that the end result will be emergent properties. Emergent properties, such as the ability to program software that can beat grandmaster-level Go players, is incalculably more impressive when you realize no one programmed DeepMind to do it.

Traditional AI is narrow and can only do what it is programmed to know, but Olli, an automated car powered by Watson, learns from monitoring and interacting with passengers. Each time a new passenger requests a recommendation or destination, it stores this information for use with the next person. New sensors are constantly added, and the vehicle (like a human driver) continuously becomes more intelligent as it does its job.

But will these AI systems be able to do what companies like Google really want them to do, like predict the buying habits of end users better than existing recommendation software? Or optimizing supply chain transactions dynamically by relating patterns from the past? That’s where the real money is, and it’s a significantly more complex problem than playing games, driving and completing repetitive tasks.

The current proof points from various AI platforms — like finding fashion mistakes or predicting health problems — clearly indicate that AI is expanding, and these more complicated tasks will become a reality in the near-term horizon.

Soon, AI will be able to mimic complex human decision-making processes, such as giving investment advice or providing prescriptions to patients. In fact, with continuous improvement in true learning, first-tier support positions and more dangerous jobs (such as truck driving) will be completely taken over by robotics, leading to a new Industrial Revolution where humans will be freed up to solve problems instead of doing repetitious business processes.

The price of not investing in AI

The benefits and risks of investment are nebulous, uncertain and a matter for speculation. The one known risk common to all things new in business is uncertainty itself. So the risks mainly come in the form of making a bad investment, which is nothing new to the world of finance.

So as with all things strange and new, the prevailing wisdom is that the risk of being left behind is far greater, and far grimmer, than the benefits of playing it safe.

First appeared at TC