Asia Summit—Artificial Intelligence: Blurring the Lines Between Humans and Machines
Robots with Empathy? They’re Coming Sooner than You Think
With present-day applications ranging from financial forecasting to elderly home care, artificial intelligence, or AI, is no longer just a science-fiction concept. Leaders representing academia, start-ups, and the corporate world shared their insights into the potential of AI at the session “Artificial Intelligence: Blurring the Lines Between Humans and Machines” during the Milken Institute Asia Summit 2016.
Despite AI’s recent popularity, there was unanimous agreement with the assessment of Pascale Fung, professor of electronic and computer engineering at the Hong Kong University of Science and Technology, there was “not a quantum jump” in AI advancement. Instead, progress has come over the past few decades as advances in computer hardware and software, algorithms and data analysis have contributed piece-by-piece to the development of AI.
Pascale Fung, professor of electronic and computer engineering, Hong Kong University of Science and Technology
Today, AI technologies are used on a daily basis in many industries. James Cham, partner at Bloomberg Beta and moderator of the session, highlighted the current uses of narrow AI, which are systems designed for specialized functions such as Siri, the digital assistant on Apple devices, or self-driving cars. For Ben Goertzel, chief scientist of both Hanson Robotics and of Aidyia Ltd., this type of industry-specific AI is already a “done deal,” particularly in the finance world.
Goertzel’s artificially intelligent hedge fund, Aidyia Ltd., illustrates his point. The use of computer models is not a novel practice in trading, but Aidyia and other artificially intelligent funds have taken it to a higher level. Aidyia uses a combination of machine learning, natural language processing, deep learning, and probabilistic logic to trade stocks and make financial forecasts weeks to months in advance, automatically improving its decision-making with each iteration.
Narrow AI has also entered the social sphere, and there are few examples in Asia more impressive than Xiaoice, a natural language chat-bot developed by Microsoft and released first in China. According to Hsiao-Wuen Hon, chairman of Asia-Pacific Research and Development at Microsoft, Xiaoice and its users have established an incredibly close relationship. The average user, who would likely participate in only two or three exchanges per chat session with a human, participates in 23 exchanges per session with the chat-bot. “A quarter of uses say ‘I love you’ to Xiaoice,” said Hon. “When we first released [Xiaoice in China], we never realized it would become this popular.”
Ben Goertzel, chief scientist of Hanson Robotics; chief scientist of Aidyia Ltd.
As vertical, market-specific narrow AI grow in capability, we will quickly find ourselves at “the dawn of the era of artificial general intelligence,” said Goertzel. Distinct from narrow AI, artificial general intelligence, or AGI, can reason, learn and perform as humans can. An AGI device would be, in effect, the stereotypical robot of science fiction. In Asia in particular, there is an opportunity for AGI to play a role in child care and elderly care — especially among Asia’s large aging population — given that humanoid robots can achieve the requisite emotional intelligence and trustworthiness.
However, the experts disagreed on whether we will see this explosion of AGI in our lifetimes. Goertzel and Fung said that within the next couple decades, humanoid robots will be a part of our day-to-day experience. Goertzel predicted that AI systems will perform all financial forecasting and risk analysis within the decade and that within two decades, the majority of current finance jobs will no longer exist. For Hon, the progression from narrow, functionality-specific AI to AGI that can truly generalize and synthesize as humans do will take longer. For the foreseeable future, the majority of people need not fear being replaced by a robot at work, he said.
Despite their disagreements on exactly when AGI will expand, all three agree that it will happen in Asia. Why does Asia stand out in its willingness to adopt AI? Hon addressed two factors that make Asian users the ideal early adopters: they are “pro-technology” and can thus tolerate imperfections as new technologies improve, and Asian users see value in using technology for social purposes. For AI in particular, Goertzel pointed to a culturally different “default mindset” in Asia about AI. People in Asia, more so than in the U.S., have believed for many years in the power of robots to help society.
These AI experts, more than the average person, can envision a world where machines learn from their experiences and comprehend the world as people do. What’s more, they can see how the algorithms and technologies we have now will lead us there. AI in today’s world, in Goertzel’s opinion, is where the personal computer was in the 1980s. “We are right at the beginning of a humongous expansion.”