Yuval Noah Harari's Concerns About AI: A Summary of His Tokyo Speech
This article summarizes a speech given by historian Yuval Noah Harari at Qingying Art University in Tokyo, focusing on his updated views on global threats, particularly regarding Artificial Intelligence (AI). The speech, hosted by Qingying Art University President Yitong Gongping, highlights Harari's evolving perspective on the risks and opportunities presented by AI.
Shifting Priorities: AI as the Primary Threat
From Triad to AI Focus
Harari, author of Sapiens: A Brief History of Humankind, previously identified three major global threats: nuclear war, uncontrolled biological technology, and the lack of control of information technology networks and AI. In this lecture, he argued that the risk level associated with AI has significantly increased compared to nuclear and biological threats. He believes the industry conversation heavily emphasizes AI's potential, while downplaying the dangers.
AI vs. Biotechnology: Speed of Development
Yitong Gongping inquired about the reason for this shift in focus. Harari explained that while both AI and biotechnology can bring significant changes, the pace of AI development is exponentially faster. Biological advancements, particularly those related to humans, require extensive periods for observation and evaluation (20-40 years). AI, on the other hand, evolves at a digital speed, potentially changing in days.
AI vs. Nuclear Threats: Potential and Perception
Harari contends that AI presents a more complex challenge than nuclear war for two main reasons. First, nuclear war has no positive outcome, creating a global taboo against aggression. Second, AI has immense positive potential, making it difficult to fully grasp the potential risks. The dangers of nuclear technology are well understood due to events like Hiroshima. The danger of AI is harder to grasp because it is a different type of threat.
The Heterogeneous Nature of AI
Not Just a Tool, But an Agent
Harari emphasizes that the core problem with AI is its heterogeneity – it's a different kind of intelligence, not simply an advanced form of human intelligence. It will surpass human intelligence but will remain fundamentally different, making its trajectory and consequences unpredictable. AI is not just a tool but an agent with mobility, capable of making decisions and inventing new things independently.
Decision-Making Capabilities
Unlike previous technologies, where humans retained complete control, AI can make decisions autonomously. This includes AI weapons that can choose their own targets and even develop new strategies without human intervention. This dual potential for good and bad is why Harari is prioritizing the history of information technology in his work.
A Decade of Change: AI's Accelerating Development
From Distant Future to Present Reality
Harari noted that ten years ago, AI was primarily a theoretical concern, a topic for science fiction. Today, its development is rapidly accelerating. While futurists like Ray Kurzweil predicted general AI by 2029, what seemed exaggerated then now appears conservative.
The Paradox of Trust
Harari points out a "paradox of trust" within the AI revolution. Leaders in AI development express concerns about the technology's potential dangers and say that they are willing to slow down. However, they argue that slowing down is impossible because competitors will continue developing faster, potentially leading to a takeover by those with less ethical considerations. Then, these leaders express trust in the super-intelligent AI they are developing, despite acknowledging their distrust of other humans.
AI as a Wave of Immigrants
Harari presents a compelling analogy: AI is not just a computer but a "global wave of immigrants," potentially displacing jobs, altering societal structures, and even taking over countries. He views this large-scale change as a significant cause for concern.
Building Trust in the Age of AI
The Trust Gap: Elites vs. Ordinary People
Ito highlights the widening "trust gap" between elites and ordinary people, questioning the reach of Harari's ideas. Harari argues that the distinction between elites and the people is an inaccurate and often misused term. He believes that the problem is not the existence of elites, but whether they are service-oriented or self-serving.
The Role of Education
The conversation shifts to the role of education. Harari believes that universities should focus on cultivating service-oriented elites. Educational institutions must move beyond simply providing information and instead teach individuals how to distinguish reliable information from the vast sea of misinformation. Students should learn to critically evaluate sources and understand the complexities of truth.
Rebuilding Trust and Addressing Global Challenges
Establishing Trust Between Humans
Harari believes the most urgent issue is establishing trust between humans, as the current lack of trust makes AI development so dangerous. Solving this problem requires multidisciplinary approaches, including biology, psychology, economics, and computer science.
Collapsing International Order
Ito expresses concerns about the collapsing international order established after World War II, citing the increasing disregard for the principle that strong countries cannot invade and conquer weak ones. This shift could lead to a decrease in medical and education budgets and a lack of AI monitoring.
Freedom of Speech vs. Freedom of Information
Differentiating Rights
In the Q&A section, Harari makes a critical distinction between freedom of speech (a fundamental human right) and freedom of information (a right being extended to algorithms and AI). He says that social media platforms intentionally confuse these two rights. The problem lies not in humans fabricating lies, but in algorithms deliberately spreading misinformation for profit. Social media companies should be held accountable for ensuring the authenticity of the information they disseminate, similar to newspapers and TV stations.
AI's Impact on Warfare
Changing the Nature of Conflict
Harari discusses how AI is changing the nature of war. While the essence of war remains killing, AI is altering who chooses the targets and makes critical decisions. AI can analyze vast amounts of data and identify targets much faster than humans, but the question is whether humans can trust AI's judgment and whether there are safeguards in place.
The Need for Caution
Harari concludes by emphasizing the need for balance in the AI conversation. While AI offers immense potential, it's crucial to acknowledge and address the potential dangers. The key is not to stop AI development, but to develop it safely, which requires rebuilding trust between humans. He reiterates that trust is fundamental to life, even down to the simple act of breathing.
{# 
{# 
{# 
{# 
{#