IN the showdown between man and machine, Google’s AlphaGo beat South Korean grandmaster Lee Sedol four games to one recently, surprising both computer jocks and professional Go players.

“I have no doubt now that AlphaGo can play Go,” declared Nie Weiping, one of China’s most famous professional Go players.

The machine shot down the prevailing wisdom that artificial intelligence was a long way from being capable to beat a human at the complicated board game of Go. That has caused a new flurry of debate in the tech realm.

AlphaGo’s victory showed the depth of artificial intelligence, popularly called AI, in terms of the development of deep learning and real-time data analysis and processing abilities. The technology is expected to have wide applications in areas such as education, healthcare, manufacturing, security and social networking. The theme was explored in the 2013 Spike Jonze sci-fi film “Her,” where an AI operating system called Samantha bonds with a lonely, depressed man.

On the other side, some people and regulators worry about the risks of artificial intelligence spinning out of control. They see the technology being applied to stock markets, creating a chasm between ordinary investors and intelligent robots. They see artificial intelligence changing the dynamics of warfare and, perhaps, some day overwhelming humans to control the world. That theme, too, has been explored in sci-fi films like “Matrix” and “Ex Machina.”

AI in daily life

Very few people realize that artificial intelligence is already part of our lives, from applications used in navigation, Siri and Google Translate to robots deployed in manufacturing and stock-advice software. AI technologies have been evolving steadily alongside the development of the Internet, big data and graphic chips.

Top IT firms like Google, IBM, Facebook, Apple, Microsoft, Baidu, Alibaba and Xiaomi are investing heavily in the future of artificial intelligence.

Jen-Hsun Huang, chief executive of graphic chip firm Nvidia, said AI has been on the stage “from research concepts to engineering applications” years ahead of public awareness.

What was once science fiction may well become stark reality as information technology converges with neurological science and intelligent robots, said Huai Jinpeng, vice minister of China’s Ministry of Industry and Information Technology.

Google announced last week that it will open connection ports of its AI platform to developers, expanding the availability of artificial intelligence to more devices and services.

“Machine learning is Next” was the topic of Eric Schmidt, the man who helped build Google from a simple search engine into one of the world’s most influential companies. He was speaking at Google’s Next 2016 conference in San Francisco last week.

Talking about endless possibilities, he said, “It’s a great time to be in the cloud.”

What was once heads in the cloud have become feet on the ground. Industry specialists see AI applications in voice recognition, image classification, machine learning, wearable computing and self-driving cars.

With the integration of neurological science and human thought patterns, devices like AlphaGo can be programmed with deep learning and self-evolutionary abilities. They can assume the work people do, often doing it better.

AI in the automotive industry

Da ze jian ji tian xia, qiong ze du shan qi sheng: In success, one strives for the greater good, while in hardship, one looks for self-improvement.

This wisdom from the ancient Chinese philosopher Mencius might describe two ways of looking at research on self-driving cars.

The quest breaks down into two camps: those seeking to make vehicles survivors, vigilant and responsive and those who want to make them peacemakers, communicative and considerate.

It all comes down to what kind of creators we want to be in applying artificial intelligence to our mobility systems.

For the makers of cars that don’t need humans as driver, it is a question of following either a pragmatic or an idealistic road. But business is about making money, not fostering ideals. Most of autonomous cars being discussed today, such as models envisioned by Google, are the stuff of pragmatists.

AI is slowly taking over control of cars, turning drivers into passengers. The change is one of evolution, not revolution. Cruise control, now a standard feature for many mid-range models, is a small step forward for artificial intelligence.

It sets a car at a certain speed so that the driver can take his foot off the pedal and cruise along on open roads free of traffic. The next step was adapting that technology to operate a car at a safe distance from the car in front and even to follow lane marking with automatic steering. Then came a car that can react to traffic lights, road signs and turn signals. These advanced driver assistance functions, which might be called “semi-autonomous driving,” are already available in high-end cars like the Tesla Model S and Volvo XC90.

The learning curve of these functions follows the improvement of algorithms — logic based on data collected from an increasing number of embedded sensors, radars and cameras. Through passive detection, which is like human observation, cars are being created to make sense of the world and be street-smart. Finding its way around relies heavily on a car’s own intelligence, with no need for other cars on the road to be equally smart. However, that “lone wolf” characteristic may make these cars vulnerable in complicated situations beyond their knowledge.

For example, blurred lane markings can be confusing to cameras, and a big truck passing by with a teetering load of goods about to spill raises no red flag to sensors. What if a ball suddenly rolls into the middle of the street? Is it too trivial for an emergency stop, or might a child suddenly dash into the road to retrieve it?

Just as humans mature with experience, artificial intelligence improves as it is exposed to more real-world happenings. That’s hard to program in a test lab.

Both Chinese and foreign carmakers are trying to simulate the real-world environment in testing V2X communications, which put cars and infrastructure on a dedicated channel for talking. This active approach is considered a step forward from passive detection-based autonomous driving because it shortens the time a car takes to react in dangerous situations and can organize vehicles to form a smooth traffic flow without unnecessary acceleration or braking.

V2X communications generate the vision of a self-controlled transportation network beyond the coordination ability of individual drivers. The challenges loom large. It is a technology that makes sense only with wide application. The creation of a shared communications platform and the investment to equip all infrastructure with communications modules are realms left to governments.

But governments are usually behind the curve where new technologies are concerned. The fact that self-driving prototype cars are still banned from most of the streets in the world limits the ability to test their capability and future.

Google and Volvo were lucky to receive permission for real-life testing from the US and Swedish government. The companies are now allowed to operate fully autonomous cars in certain areas.

Current laws and regulations stipulate that auto functions may assist drivers but not replace them. That means those who are among the first to drive these cars may be unwilling to fully use the advanced functions.

A new threat?

Artificial intelligence both awes and scares us. Red flags appear. Experts in the field say more evaluation needs to be done on how AI is used.

It’s difficult to assess the implications of artificial intelligence if we don’t fully understand how machines make decisions. Can regulators really control the use of AI in market manipulation, advertising invasion of privacy and illegal web searches?

Then there’s “singularity” time, expected around 2045, when artificial intelligences is forecast to play the major role in our lives, doing things that people can’t do.

The question remains: will AI be a benefactor to humankind or will it become the Frankenstein monster of the 21st century?