What Happens When Digital Superintelligence Arrives
The evolution of artificial intelligence (AI) is already moving beyond the stage of "mimicking humans." Having passed through generative AI and artificial general intelligence (AGI), the focus of discussion has now shifted to "Artificial Superintelligence (ASI)."
In this article, based on a conversation between Dr. Fei-Fei Li and Dr. Eric Schmidt at the FII (Future Investment Initiative) we will organize the definition of superintelligence, the limits of creativity, the impact on the economy, and the role of humanity.
1. What is Superintelligence: The Arrival of Intelligence That Surpasses All Humanity
Dr. Schmidt defines superintelligence as the sum total of all human intelligence, or something that exceeds it. According to him, some AI researchers, known as the "San Francisco Consensus," predict that
superintelligence will arrive within 3 to 4 years, but he himself takes a cautious stance, saying it will take longer than that.
While acknowledging that AI has already surpassed humans in specific domains, Dr. Li emphasizes that whether AI can make creative discoveries like Newton, Einstein, or Picasso is a different matter. She points out that "even if you give AI all the astronomical data, it cannot derive the law of universal gravitation on its own." This indicates that
there is a deep chasm between data processing capability and creative abstraction.
2. Technical Challenges: The Wall of Creativity and Algorithms
Dr. Schmidt states that new algorithmic breakthroughs are essential for achieving superintelligence. Current AI is designed to optimize for fixed objective functions and does not possess the "creativity to dynamically redefine goals themselves" that Dr. Li speaks of. To overcome this, he says, AI will require non-stationarity of objectives, where it can reset its own goals.
While citing the example of GPT-5 Pro reaching an IQ of 148, Dr. Schmidt sounds an alarm, saying, "We are entering a realm where the 'meaning of intelligence' beyond the upper limit of IQ is being lost."
3. Impact on Economy and Society: The Uneven Distribution and Redistribution of Wealth
AI is projected to generate 15 trillion dollars in economic value by 2030. However, Dr. Schmidt points out that "the expansion of wealth does not necessarily lead to shared prosperity." While efficiency gains from AI will certainly create wealth, there is a danger that it will
concentrate in a small number of companies and nations due to network effects.
Dr. Li agrees, but emphasizes that while "AI will democratize knowledge, healthcare, and transportation, shared prosperity is a matter of policy and social structure, not technology." The "democratization of capabilities" brought about by AI is inevitable, but the "redistribution of wealth" remains firmly in the realm of politics.
4. National Strategy and Geopolitics: Whoever Controls AI Infrastructure Controls the World
Dr. Schmidt asserts that the birth of superintelligence will be left in the hands of hyperscalers. The United States holds an overwhelming advantage backed by chip manufacturing like TSMC and massive capital markets, with China in pursuit. Meanwhile, countries like Saudi Arabia and the UAE are building their AI foundations through partnerships with US companies.
Dr. Li offers a realistic view, stating, "Not every country should have its own data centers." Small countries are at a cost disadvantage, but she proposes
establishing AI sovereignty through regional cooperation and joint investment. Concerned about the lag in African nations, she also stated that "human resource development and the construction of stable industrial structures are urgent tasks."
5. The Role of Humanity: Protecting Creativity, Ethics, and Dignity
Finally, the discussion turned to human irreplaceability in the AI era. Dr. Schmidt says, "Even if robots surpass humans in the future, we will continue to love competition and art among humans." Dr. Li emphasized that
we must place human dignity and agency at the center.
"The essential purpose of technology and policy is to place human well-being at the center. No matter how much AI evolves, we must not choose a future that excludes humans."
These words can be called an ethical compass in an era swaying between the frenzy and fear of superintelligence.
Conclusion: Toward Co-evolution with Superintelligence
The possibility of AI surpassing human intelligence is no longer a fantasy. However, for it to be a
future that evolves alongside humanity rather than a "future that surpasses humanity," the four pillars of creativity, ethics, policy, and education are indispensable.
The dialogue between Dr. Fei-Fei Li and Dr. Eric Schmidt confronts us with the question,
“The question is not the speed of technological progress, but the speed of social maturity”.