(Image source: pixabay)

Hollywood has made countless AI-related movies using different techniques, and AI has almost become a subgenre of Hollywood movies. In the 1973 movie Wesorld, Yul Brenner played an AI tough guy cowboy robot who walks into a bar and gets into a shootout. A few years later, Disney had a very different take. In Big Hero 6, Baymax, a giant robot as soft as a pillow, lovingly accompanies his 14-year-old master on a tense *** journey. "He'll change your world," the movie says.

Yes, AI will change our world and augment human intelligence, and the role of AI in assisting humans is more like Baymax than Berlena's cowboy robot.

Big data, massive computing power, and complex algorithms are the three breakthroughs that are accelerating AI from sci-fi movie to reality. Thanks to the exponential growth in the number of cameras and sensors in our daily lives, a huge amount of information is being collected and utilized at an astonishing rate. Artificial intelligence needs to learn from information. The cloud has given everyone access to huge amounts of computing power, and we now have sophisticated algorithms that can pull out the highlights and distill insights and wisdom from the mountains of data.

But the current AI is not quite the same as Cupcake or Burrina, and it's still a long way from the so-called "artificial general intelligence" (AGI), which is when computers and human intelligence go hand in hand, or even surpass humans. Like human intelligence, artificial intelligence can be categorized according to levels. At the bottom is simple pattern recognition. In the middle is perception, which senses increasingly complex scenarios. According to estimates, 99% of human perception is through language and vision. Finally, the highest level of intelligence is cognition, the ability to y understand human language.

These are the foundations of artificial intelligence, and Microsoft has invested heavily in research and development at these levels over the years. For example, there are statistical machine learning tools that can make sense of data and recognize patterns, and computers that can watch, listen, move, and even begin to learn and understand human language. Under the leadership of our Chief Computer Language Scientist, Xuedong Huang, and his team, Microsoft set a new record for correctness, with our computer systems transcribing telephone conversations more correctly than trained professionals. In computer vision and learning, in late 2015, our AI team won five major challenges, even though we had only trained the system to solve one of them. In the Common Objects in Context challenge, the AI system had to try to solve several visual recognition tasks.

We've only trained our system to do the first challenge: look at a photo and label what you see. However, with a little transfer learning, the neural network we built was able to learn and autonomously complete the other challenges by not only interpreting the photo, but also circling each individual object in the photo and describing the action taking place in the photo with a single sentence of text.

I believe that within a decade, AI will be able to outperform humans in speech and visual recognition. But just because a machine can see and hear doesn't mean it can learn and understand. Natural language understanding, or human-computer interaction, is the next new frontier.

So in the future, can artificial intelligence really reach the realm that everyone imagines?

So in the future, will AI really be able to reach the level that we all think it will, and how will it benefit everyone? Again, the answers are hierarchical.

Tailoring. Today, we're at the very bottom of the artificial intelligence spectrum. Artificial intelligence is customized and tailored. A handful of tech companies with access to data, computing power, and algorithms build AI products and bring them to market. Only a handful of people can build AI for the majority. This is where most of the AI is now.

Universalization. The next stage is universalization. Microsoft, as a platform company, has a long history of providing the underlying technology and tools on which others can build to develop more innovation. Our strategy is to put the tools to develop AI in everyone's hands. Promoting AI means enabling every person and every organization to create unexpected AI solutions that meet specific needs. We can use the analogy of the popularization of movable type versus the printing press.

In the 1450s, there were only about 30,000 books estimated to exist in Europe, each one handmade by someone in a monastery. The Gutenberg Bible was the first book to be produced using movable type, and over the next 50 years, the number of books grew to about 12 million, leading to a renaissance in learning, science, and the arts.

Artificial intelligence needs the same trajectory. To get there, we need to be inclusive and democratic. So our goal is to make tools with artificial intelligence that are integrated into all personal assistants, apps, services, and infrastructure:

◆ We are working to use artificial intelligence to comprehensively change the way humans interact with personal assistants such as Cortana, which is becoming more and more prevalent in our lives.

◆ Apps like Office 365 and Dynamics 365 will incorporate artificial intelligence to help us focus on what matters most and make the most efficient use of our time.

◆ We will open up several of Microsoft's essential intelligence services, including pattern recognition, perception, and cognitive capabilities, to app designers around the world.

◆ Finally, we are building the world's most powerful AI supercomputer and working toward opening up this infrastructure to everyone.

Many industries are beginning to utilize AI. McDonald's is developing an AI ordering system to help employees at drive-throughs make ordering easier, more efficient, and more accurate, Uber is using Microsoft's Cognitive Services tools to compare driver photos and identify drivers to prevent fraud and improve passenger safety, and Volvo is using our AI tools to detect driver distraction and can alert drivers to prevent accidents.

If you're a business owner or manager, you can imagine having an AI system that can see all of your company's operational processes, understand everything that's going on, and notify you of the things that matter most. Image analytics company Pri *** Skylabs innovates on the basis of Microsoft's cognitive services, allowing computers to manipulate image surveillance cameras and analyze current events. If you run a construction company, an AI system can notify you that a cement truck has arrived at a site. For a retailer, the system could track inventory or assist you in locating the store manager of a particular location. One day in the future, the system might be able to observe surgeons and assistants in hospitals and, if it detects medical negligence, warn the medical team before it becomes irreversible.

Learn how to learn. The ultimate artificial intelligence is when computers learn how to learn, that is, when they can generate their own programs. Like humans, computers will eventually be able to do more than just mimic human behavior; they will be able to create newer and better ways to solve problems. Deep neural neorks and transfer learning have given us breakthroughs, but artificial intelligence is like a ladder, and we're still at the bottom. At the top of the ladder is strong artificial intelligence and a machine's full understanding of human language. In time, computers will be able to demonstrate intelligence that is on par with humans, or indistinguishable.