On Saturday July 27, the Jintai Art Museum provided a beautiful venue for Humanity+@Beijing, the first conference in Beijing to focus on the implications of radical technological advance for the future of humanity. Over 60 attendees joined a roster of Chinese and Western speakers for a wide-ranging exploration of the future.
Following Humanity+ @ HongKong in 2011, Humanity+ @ Beijing was the second Humanity+ conference in Asia. Humanity+ @ Beijing was organized by Humanity+ in collaboration with Beijing Commons, an organization that has organized a variety of technology-oriented conferences in Beijing including BarCamp and many others.
While touching a variety of themes, the central focus of Humanity+@Beijing was AI and robotics, and the future implications of these technologies for economy and society, and humanity in general. A number of the speakers and attendees were in Beijing for other AI-related events happening in early August: the AGI Summer School and AGI-13 conference at Peking University (http://agi-conf.org.2013) and the International Joint Conference on AI (http://ijcai13.org).
Adam Ford, an Australian Humanity+ Board member who has organized multiple Humanity+ and Singularity focused conferences in Melbourne, began the conference with a broad overview of the concept of the Intelligence Explosion: the point at which machines begin improving their own intelligence, hence causing a very rapid increase in their intelligence level, going far beyond human comprehension.
His fellow Australian, the entrepreneur James Newton-Thomas, brought the same themes back to Earth, discussing work by himself and others applying AI to the control of various complex machines, especially in the mining industry. Topics ranged the use of partially autopiloted drones for search and rescue, to the use of AI to control mining equipment and potentially whole mines… These achievements were presented not as isolated innovations, but rather as part of an overall movement toward automation of difficult or dangerous physical labor traditionally done by humans.
Economist Jack Strocchi, recently retired from a post with the Australian government, continued Newton-Thomas’s direction with a discussion of the economic implications of the displacement of human labor by increasingly intelligent robots. He expressed the view that Asia’s economies, including China’s might prove better equipped to handle these changes than those of the West, due to their hybridization of state and corporate control of major industries. It may be, he suggested, that a combination of collectivist and profit-motive based methods are needed to adapt society and economy to a future in which robots do all the work. Ultimately, though, Strocchi’s conclusion was somewhat dour, positing a future world in which humans have no economic purpose, yet maintaining the value of humanity in asserting moral and aesthetic judgment, areas in which he believes humans intrinsically more capable than machines.
The morning session was capped off with an inspirational talk by multidisciplinary creative director/designer and Humanity+ Board member Amy Li, who discussed the importance of creating technology that harmonizes with human emotion. Rather than robots that take human jobs, Amy focused on a future consisting of technology that interweaves with every aspect of human life, responding to human emotion and helping people to experience life more satisfyingly and more deeply.
As a graduate from Singularity University, Amy shared her insider experience at SU’s Graduate Studies Program last year, and her partly SU-inspired entrepreneurial journey in the area of biosensors and “quantified self” type data gathering. This journey eventually led her to Dance4Healing, a project using behavior design and AI to build a ecosystem aimed at unleashing the healing power of music and dance/movement therapy, starting with cancer patients. Amy illustrated the ideas behind Dance4Healing by teaching the audience two simple hand dances that are scientifically proven to provide neurological benefits when practiced regularly.
The afternoon session began with Dr Hugo de Garis, a physicist, computer scientist and futurologist famed as the inventor of evolvable hardware. DeGaris recently retired from Xiamen University but is spending his “after retirement career” actively pursuing both research and education in advanced mathematics and physics. He addressed the need for change in the publishing industry, given the capability the Internet provides for broad and low-cost distribution of knowledge. He predicted that paper publishing companies will die off within a decade, replaced by a methodology of “globacation” or global education, in which information is provided by the people of the world for the people of the world, directly via the Net without the need for the mediation of publishing companies. Indeed there is evidence this is already happening, with online courses like Coursera and Udacity, as well as Dr. de Garis’s own online math and physics courses and the tremendous amount of educational and research material available online. De Garis also predicted that, with the rise of global education, the boundaries between nations will become increasingly irrelevant and the world will converge into a global state. On the darker side, he also reiterated his prediction that the 21st century will be marked by an Artilect War, a violent conflagration between proponents and opponents of advanced technologies. Other speakers and conference participants appeared to find de Garis’s predictions interesting yet not entirely compelling.
The rest of the afternoon focused on the practicalities of achieving advanced artificial intelligence and highly capable robots. Immediately following de Garis was a talk by my own self, Dr. Ben Goertzel (in person) and Dr. David Hanson (appearing via Skype), on our work applying the OpenCog Artificial General Intelligence architecture to control intelligent agents. Video footage was shown from Raj Dye’s award-winning documentary film “Singularity or Bust!”, which featured an Aldebaran Nao humanoid robot controlled by software created by de Garis and myself and our collaborators, filmed at Xiamen University in 2009. The robot carried out conversations about its immediate environment and philosophical topics, as well, using a natural language dialogue system engineered by Xiamen University PhD student Ruiting Lian.
A series of videos was shown, highlighting Hanson’s remarkably realistic-looking humanoid robots, ranging from Robot Einstein to a female robot singing emotional music, and meter-tall robot toddlers that are the focus of Hanson’s and my collaborative research. These robots feature a novel nanotech-based skin called Frubber, which allows unprecedentedly realistic humanlike facial expressions. Hanson appeared via Skype, together with the robotic head of science fiction writer Philip K. Dick, and the audience initially could not tell which head belonged to the speaker and which to the robot!
According to the plan I articulated in my talk, creating a robot toddler is the first step toward realizing the grander goals discussed by Newton-Thomas, Ford and the other morning speakers. A robot toddler would represent the achievement of AI able to understand human common sense knowledge about the everyday world; and then based on this foundation, the same AI system would then be able to achieve all manner of practical wonders. The potential of future AI-driven innovation in medical research, energy development and space exploration was also highlighted. Reference was given to the OpenCog project website at opencog.org, and the GENI Lab website (geni-lab.com) hosting Hanson’s and my robotics collaboration, which also includes Mark Tilden, the Hong Kong-based creator of the Robokind humanoid robots, with over 23 million sold, and our Hong Kong Poly U collaborator, consciousness researcher Dr. Gino Yu. The four of us are currently seeking donations to further this work via http://geni-lab.com/help .
The final two lectures were by young AI researchers in town for the AGI Summer School at Peking University. Yiqing Liang discussed the potential of a kind of AI called “deep learning” to emulate the way the human brain recognizes objects and events in its visual field, and discussed her entrepreneurial work in the UK aimed at applying deep learning to help automatically diagnose human diseases. Cosmo Harrigan reviewed in detail the reasons why embodying AI systems in virtual worlds, such as 3D game worlds, may be a powerful modality for refining AI methodologies and teaching young AI systems. Even if one’s end goal is robots that relieve humans from repetitive or dangerous labor, or explore space, AI agents inhabiting virtual worlds may be a natural incremental step along the path.
The audience had numerous questions and comments throughout, both about the practicalities of advanced AI and robot emotion, and the long-term future of human civilization given the prospect of such radically transformative technologies. Given the vibrant discussions and positive reception, future Humanity+ @ Beijing events are likely.