Do you remember Arnold Schwarzenegger’s signature movie The Terminator?
I watched my all-time-favorite again recently. One of the scenes shows how Arnold, the robotic killer, welded and fixed the damaged parts of his body in a shabby motel room. A cleaner passing by smelled the stink and asked if there was any dead cat inside. A few options for the response popped up for Arnold: “yes” / “ no” / “please come back later” / “piss off!”. The intelligent robot picked the last answer which best suited the situation.
I am still fascinated by the artificial intelligence (AI) model shown in the 1984 production, which still looks futuristic even today. Now, 32 years after the movie, we are still developing and refining AI!
AI is about to take over the world
AI is generally defined as “the study and design of an intelligent agent which is a system that perceives its environment and takes action” that maximizes its chances of achieving its objectives.
It has been widely used in various areas. A few years ago, Siri which could respond to your questions like a living human being was the favorite toy of many iPhone fans.
Wikipedia, with almost all of its articles available for editing by anyone and registering 10 changes made globally every 10 seconds, started to deploy an AI robot to address accusations of inaccuracy. The robot is able to detect bad or malicious edits and alert Wikipedia’s human editor to take action.
Local AI development
In Hong Kong, the development of AI has also been picking up in recent years. The City University of Hong Kong was awarded the Best Business Application Grand Award by Hong Kong ICT Award in 2014 for its Integrated System for Engineering Works Management built for MTRC. According to the judging panel, the system innovatively deployed AI technologies to manage numerous combinations of data including manpower, wagon, locomotive, equipment, location, and train speed. It aims to provide conflict checking and work optimization for more than 2,600 engineering works every week. By handling the complexity, the system is able to help MTRC bring safety to its more than five million passengers every day.
Another example is the use of AI for stroke diagnosis. The Hong Kong Polytechnic University has developed a new application that “expertly analyses brain scans” to help doctors determine if a patient is in the life-threatening condition. Immediate diagnosis and treatment within the first three hours of ischemic stroke can greatly minimize brain damage. Therefore, the application which takes less than 10 minutes can help save lives.
Late last year, The Hong Kong University of Science and Technology launched the WeChat-HKUST Joint Laboratory on Artificial Intelligence Technology (or WHAT LAB). The mission of the lab is “to foster artificial intelligence and big data research to improve people’s lives and advance the frontiers of knowledge.”
For WeChat, which is holding a vast amount of data from its 650 million active users, the integration of AI and data mining technologies can reveal more knowledge and insights about users’ preferences and habits. This will enable WeChat to serve its customers better.
Challenges for AI
Straight-forward and toilsome tasks are exactly what AI is good at. However, complicated tasks leading to an ambivalent attitude similar to the response in The Terminator movie pose a great challenge for AI designers to imitate the human thinking process.
Last November, a motorcycle cop in the US stopped a driver-less car and warned the two engineers on board for creating a hazard, as it was going at 24 miles an hour within a busy 35 mph zone. The cutting-edge, law-abiding creation is supposed to lead to a world free of accidents. But it is now doing the exact opposite unfortunately: the AI operated vehicles have racked up a crash rate double that of human drivers, according to a study by the University of Michigan. These were all minor scrape-ups most of which are ironically caused by inattentive human drivers.
So the major challenge for AI is how to integrate the human behavior and decision making process, which can be intuitive, illogical, even chaotic, in addition to the normal conscious step-by-step deduction.
Researchers at Stanford University in the US have a creative approach on this issue: by using crowdteaching. They developed a virtual 3D driving game with California’s highways as the background, and invited human drivers to play the game. For the first time, an AI machine can learn complex driving skills from the behavior of real people.
According to media reports late last month, China’s Ministry of Industry and Information Technology is about to announce the next five-year (2016-2020) implementation plan and one of the focus areas is artificial intelligence.
Computers taking over the world, or not?
So will AI take over our world one day?
A well-known critic of AI, also the co-founder of Paypal and Tesla, Elon Musk, does not think so. He recently formed a non-profit venture called OpenAI. The goal of the venture is “to ensure that the scary prospect of computers surpassing human intelligence may not be the dystopia that some people fear.”
The new organization will pursue the most advanced forms of AI and make its findings known on a royalty-free basis to the public.
By the use of AI, we can avoid repeated and mechanical works. This enables us to maximize the uniqueness of human beings – creativity and imagination – to the fullest extent.
On that note, may I suggest that youngsters in Hong Kong grasp this golden opportunity to develop a more beautiful world!
Dr. Winnie Tang
Co-founder and Chairman, steering committee of Smart City Consortium