Multi-channel : China news + Kids in HK
The question of when AI will surpass human intelligence, often referred to as achieving "superintelligence," has been the subject of intense debate among industry leaders.
In this last episode on digital superintelligence, I want to compare predictions by industry leaders regarding when artificial intelligence will outsmart humans and discuss key drivers for its rapid rise.
For example, Ray Kurzweil, a renowned futurist and AI researcher, predicts that AI will reach human-level intelligence by 2029 and that the Singularity - the point where humans merge with AI - will occur by 2045.
His bold predictions conjure up a future in which humans merge with AI through brain-computer interfaces and nanobots, enhancing our intelligence and consciousness.
Tesla and SpaceX chief executive Elon Musk predicts that AI will surpass human intelligence within the next few decades, potentially as early as 2026. His rationale is rooted in the accelerating pace of AI development and the potential for AI to mimic and exceed human cognitive abilities.
In summary, the rapid rise of AI can be attributed to five key factors that will play a critical role in determining the timeline.
Data advancement is the first driver.
Brian Gallagher, the former chief executive of United Way Worldwide, the largest privately-funded nonprofit, suggests the global dataset is doubling every two years.
The proliferation of digital devices and the internet has led to an explosion of data generation between 2010 and 2035. Data serves as the foundation for training AI models and improving their performance.
Next, techniques such as neural networks and natural language processing have revolutionized AI capabilities.
Significant progress in machine learning and deep learning algorithms has enabled AI systems to achieve unprecedented levels of accuracy and efficiency.
DeepSeek, a Chinese AI company, claims to have trained its model using less advanced chips. This has sparked discussions about the feasibility of training large AI models without requiring billions of dollars in capital and cutting-edge technology.
With two eras of computer usage in AI training, Singularity University cofounder Peter Diamandis suggests the cost of AI training has gone down 99.5 percent over the past five years.
The third driver is computing power.
Computational power doubles every two years, often referred to as Moore's Law.
The development of powerful hardware, such as graphics processing and tensor processing units, has accelerated the processing capabilities of AI systems.
This has allowed for the training of larger and more complex models, leading to breakthroughs in AI performance.
The "All-in" capital flow on AI is the fourth factor.
Last year, just three companies - Meta, Microsoft and Google - invested US$152 billion (HK$1.18 trillion) in AI and have shown no signs of slowing down.
The exponential growth of investment in AI research and development by governments, corporations and venture capitalists has fueled the growth of AI.
This financial support has enabled the scaling of AI projects and commercialization of AI applications.
Interdisciplinary collaboration is the final factor.
Open collaboration has facilitated the development of innovative AI solutions. Researchers from diverse fields bring unique perspectives and expertise, driving the advancement of AI technology.
As I pen this final article for Tech in Synergy, I want to express my heartfelt gratitude to all of you - my loyal readers.
Since May 2020, we have embarked on an incredible journey together, exploring the ever-evolving world of technology and its profound impact on our lives. Your thoughtful feedback has been the driving force for each of the 225 articles we've shared.
Due to other commitments, I must now end this column for a while. Thank you for being a part of this remarkable adventure. Until we meet again, stay curious, stay inspired and keep embracing the future.
Dr Jolly Wong is a policy fellow at the Centre for Science and Policy, University of Cambridge
