Artificial intelligence is changing the world in the 21st century just as profoundly as steam engines did in the 19th century and motor vehicles in the 20th.
The changes are all around us already, have often been seamlessly integrated into everyday life and have only just begun, Prof Yike Guo, Provost of the Hong Kong University of Science and Technology, Chair Professor of Computer Science and Engineering and leading AI researcher notes.
Online navigation, where cars instantly tell drivers the shortest way to a destination and can read out or display the route on a map, has been around for many years and is improving all the time. Prof Guo said AI use has become a natural part of our everyday world and we often deploy it without even realising.
“The last time you communicated with your bank online, was that a person or and AI programme? Almost certainly a programme,” he said.
We passed the Turing Test – a thought experiment by Second World War era computer scientist Alan Turing who theorised that if it was impossible to judge whether or not a typed conversation was machine generated, then the machine was “intelligent” – long ago almost without noticing, Prof Guo said.
And one of the leading centres of AI and the related robotics development was Hong Kong, he pointed out, naming Hanson Robotics, developer of the iconic humanoid Sophia, and facial recognition pioneer SenseTime as just two examples of world-leading AI innovators based in Hong Kong.
With five of the world’s top 100 universities, and the only city in the world with such a tally, Hong Kong is also a leader in academic AI development.
AI is difficult, theory-heavy and expensive, with vast amounts of data storage and processing required, Prof Guo emphasised, which meant “zero to one” development – going from concept to the first deployable prototype – needed to take place in universities. Many other forms of innovation took place on the factory floor or out in the market, he conceded, but AI remained primarily in universities.
AI and robotics comprise a gigantic field. Almost any human task can involve learning machines. Prof Guo has been working on art and his AI-generated portrait of the late leader Deng Xiaoping in front of Victoria Harbour has drawn a great deal of media attention.
He is team leader of a still-running project, Building Platform Technologies for Symbiotic Creativity in Hong Kong, which explores creating art that actively responds to the reactions of the audience seeing, hearing or touching it.
In addition to HKUST, the project includes researchers from Hong Kong Baptist University, City University, the Hong Kong Polytechnic University, Shanghai University, and Tsinghua, as well as several overseas universities including Imperial College London.
He has also led projects creating AI-generated choirs giving animated performances. In another project, he is leading a team producing a version of Vivaldi’s Four Seasons that can respond to the audience directly based on their reactions.
Prof Guo divides AI into two distinct fields – descriptive AI and generative AI. Descriptive involves identifying things, putting them into categories and distinguishing objects. SenseTime’s work on facial recognition is a typical example.
Generative AI, which has created much of the excitement recently, involves using AI to create things, with Chat GPT and his own work on visual and musical arts as classical examples.
AI has brought two related fears. The first is that machines will eliminate the jobs of creative and knowledge workers, much as automation and robots appear to have done to manual labourers. The other is of an apocalyptic scenario dreamed up in the Terminator and Matrix film series, in which super-intelligent machines eliminate or enslave humans.
Prof Guo dismissed the redundancy fear, saying that AI, like all tools, would make the work of humans better and easier, but not replace them. Humans who once carried loads had become drivers of horse-drawn carts and then trucks – the horse became redundant, the driver did not.
On security fears, he said AI was a powerful tool and like all tools could be misused. There was a need for tight international regulation.
“We have international regulations for accounting, but AI is far more complex than accounting,” he pointed out.
Realising that regulation would be a difficult, lengthy task requiring global cooperation, HKUST was leading an international meeting of AI experts in Hong Kong this month to discuss this very subject – regulation of AI use worldwide.
The International AI Cooperation and Governance Forum 2023, hosted by HKUST and Tsinghua University runs on 8 and 9 December and draws delegates from around the world who will address ways to counter the challenges AI has brought in areas, such as privacy and security, traceability and liability, malevolent deployment for disinformation, fraud and more. These have all been hot-button issues for information technology for many years, but AI greatly expands the potential scope for harm.
Hong Kong’s position as a global communications and education hub made it an ideal centre for AI development, Prof Guo believed, since collaboration between universities across the world was critical to development.
His own education was a good example of such diverse development. Originally from Shanghai, Prof Guo studied in Tsinghua University in Beijing as well as Imperial College London, which he described as the birthplace of logical programming.
He became interested in AI early on, while studying in Tsinghua.
Looking ahead over the next five years, Prof Guo sees AI becoming more tightly integrated into our daily lives, while the processes of the systems themselves would become closer and closer to natural human thought.
