Why Artificial Intelligence Will Dominate the World18 December 2018
Artificial Intelligence is older than many people think. The development of technologies around this concept has existed for at least six decades. The term was first used by the American mathematician John McCarthy in the 1950s. However, according to engineer SK Reddy, one of the top AI experts in Silicon Valley, the advance has been meteoric in the last five years, impacting decisively on the business model of many companies. SK Reddy is CPO of Artificial Intelligence at the Swedish multinational Hexagon, a reference in digital solutions. Recently, he was in Brazil to speak at an event on behalf of Hexagon's Agriculture division. The engineer cited three movements that have propelled the advancement of AI in recent years and should continue to fuel its development in the coming decades. They are:
- Universities like Stanford and Silicon Valley University as well as innovative companies like Google, Amazon, Facebook and Baidu are developing frameworks - software libraries and languages - which greatly facilitate the work of developers, helping them think and use Artificial Intelligence in their solutions. An example is IBM Watson, widely used to improve customer service and support services.
- The volume of data freely accessible is infinitely greater than in the recent past. Big Data analyses are fundamental for the realisation of experiments and tests, collaborating with the advance of many technologies, including the application of Artificial Intelligence.
- There has been a very large development in hardware in recent years. Less than a decade ago, we had only CPUs (central processing units), which evolved into GPUs (graphics processing units), more conducive to the international intelligence market by accepting more complex calculations. Today, Google is already focusing on the development of TPUs (tensor processing unit) to further accelerate processing in AI.
For SK Reddy, these three factors are making it easier, cheaper and faster to develop and experiment with solutions based on Artificial Intelligence technology. Among the most popular applications we can cite the ads that appear on the pages accessed by the user just after he or she has researched the same product in an online store. Or the Netflix indications based on what the subscriber has already watched on the platform. The point is that AI is a long way from restricting the automation of human tasks, such as driving a car from point A to point B.
There are endless possibilities, which according to SK Reddy will be known at incredible speed. In agriculture, for example, Artificial Intelligence is already used to automate processes and equipment such as tractors and harvesters, reflecting more accurately the sowing or application of pesticides. Soon, AI will allow automatic identification of pest infestation – anticipating ways to fight that infestation. AI will issue an alert when an agricultural machine needs maintenance. "Do not be surprised if in a few years you pick up the phone to make a call and the device already inserts the number you want to call in advance," warns SK Reddy.