Taiwan’s Academia Sinica President Liao Junzhi presented at the annual conference of the East Coast Chinese Academic Consortium on August 16th, highlighting major innovations and the rise of artificial intelligence (AI) over the past thirty years. He emphasized the importance of seizing opportunities based on understanding human nature and seeking evidence to filter out noise for further research.
Over the past thirty years, humanity has experienced significant innovations such as the internet, CCD cameras, smartphones, GPS systems, social media, electric cars, mRNA vaccines, and AI. These innovations have witnessed exponential growth, with the emergence of chatGPT in 2022 causing a significant stir in the field.
Liao Junzhi outlined the evolution of AI over several stages: from its emergence in the 1960s as a field of research within computer science, to pivotal moments such as AlphaGo defeating the Korean Go champion Lee Sedol in 2016, marking the first time AI outperformed a human world champion. Subsequently, AI underwent qualitative changes, shifting towards autonomous learning and rapid advancements surpassing human capabilities in various domains.
In 2022, the sudden development of large-scale models like ChatGPT by OpenAI, despite other tech giants like Google and Microsoft also working on similar projects, took the world by surprise. ChatGPT not only processes language but can also engage in reasoning, challenging conventional perceptions of AI limited to specific tasks like translation or writing.
AI’s rapid advancement, according to Liao Junzhi, is attributed to the combination of goal setting, screening mechanisms, and reward systems, allowing for swift evolution cycles leading to progress in a short period. Aligning goals, screening, and rewards can facilitate rapid improvement in organizations and educational settings.
While highlighting the potential of AI, Liao Junzhi acknowledged its limitations, emphasizing its reliance on existing knowledge for interpolation and optimization rather than the systematic generation of entirely new theories surpassing human cognition. He also pointed out constraints related to energy sources, particularly the need for clean energy like nuclear fusion, and computational power limitations which could be potentially overcome with quantum computing in the future.
Liao Junzhi stressed the value of human creativity in generating novel ideas and noted the unpredictability of the future. Instead of precise future predictions, he emphasized the importance of understanding trends and boundaries to avoid misjudgments. His motto “洞燭先機” encapsulates the strategy of being ahead of the curve by understanding human nature, basing decisions on logic, seeking evidence, and filtering out noise for effective research towards future developments.