Alexandr Wang (Scale AI Co-founder) – The Futurist Summit (Oct 2023)

Chapters

Abstract

Harnessing AI for Global Leadership and Ethical Challenges: Scale’s Vision and the Geopolitical Landscape

In the rapidly evolving field of artificial intelligence (AI), Scale, founded by ex-MIT AI researcher Alexandr Wang in 2016, has emerged as a pivotal player. With its roots in AI infrastructure and data platforms, Scale’s collaboration with giants like OpenAI, Meta, and Microsoft reflects its differentiated approach to democratizing AI access. However, the company’s involvement with the Pentagon and military AI, spurred by Wang’s concern over geopolitical dynamics, particularly China’s aggressive AI strategy, signals a complex interplay of technological prowess and ethical quandaries. This article delves into the nuances of AI’s role in military power dynamics, ethical considerations in AI deployment, and the imperative of human involvement in AI, set against the backdrop of an intensifying AI arms race between the U.S. and China.

The Vision and Core Business of Scale:

Alexandr Wang, the founder of Scale, grew up in Los Alamos, New Mexico, where his parents worked for the National Lab on fusion physics and weapons technology. This environment instilled in him a deep appreciation for the dedication of brilliant scientists working to maintain U.S. leadership in advanced technology. With this background, Wang established Scale with the ambition of empowering AI development, focusing on data and foundational technologies. The company’s core business lies in providing AI infrastructure, working with key sectors like autonomous vehicles and government agencies. Scale’s unique strategy lies in supporting the entire AI ecosystem, aiming to make AI tools universally accessible.

Scale’s Involvement with the Pentagon and Military AI:

Wang’s background and the AI advancements in China have motivated Scale’s collaboration with the U.S. government. China’s aggressive AI development strategy, aiming for global dominance, has been a significant driver for Scale’s strategic direction. China’s “Made in China 2025” strategy emphasizes chips and AI as key tenets, with the aim to dominate the future world through chips and semiconductors. Wang’s observations during a visit to a Chinese facial recognition company in 2018 reinforced his concerns. He witnessed real-time face recognition technology in action and realized that China was dedicated to using AI to power its country’s ambitions and build a global surveillance state.

Potential Risks and Negative Impacts of Advanced Chip Technology

Misuse in Cyber Attacks:

Advanced chips can launch sophisticated cyber attacks on critical infrastructure, financial systems, or sensitive data, causing widespread disruption and significant financial losses.

Misuse in Bioweaponry:

Advanced chips can be used to manipulate genetic material, enabling the creation of highly targeted and lethal bioweapons that are difficult to defend against.

Negative Consequences for Humanity:

The misuse of advanced chips in cyber attacks and bioweaponry could have far-reaching and severe consequences for humanity, posing a significant threat to global security, stability, and well-being.

The AI Arms Race with China:

The U.S. currently leads in AI innovation, but China has ambitious plans to catch up. China has demonstrated its ability to catch up and even surpass U.S. tech capabilities in certain areas. Wang fears that the same could happen in the current wave of large language models and generative AI. The escalating need for resources like high-end GPUs for AI development adds a complex layer to global geopolitical tensions. AI development relies heavily on hardware, particularly high-end GPUs manufactured by NVIDIA. Scaling laws dictate that AI systems are using dramatically more compute, data, and larger models to build more powerful algorithms. Over a seven-year span, there has been a 100,000-fold increase in computational power applied to training large generative models. The increased demand for high-end GPUs creates pressure on the supply chain, especially in light of recent export controls on chips. This issue is likely to become a contentious topic in the US versus other countries. Currently, 100% of high-end GPUs are manufactured in Taiwan, making the supply chain vulnerable to geopolitical tensions.

Geopolitical Tensions and AI’s Military Role:

Taiwan’s TSMC fabs are the choke point for AI progress, creating geopolitical tensions. In a scenario of China’s perceived technological inferiority, an invasion of Taiwan could disrupt AI advancements or lead to China’s control of TSMC’s production. AI is one of the few technologies with the potential to shift the balance of power in warfare. The Chinese Communist Party (CCP) sees AI as a potential leapfrog technology for the PLA against the U.S. DoD. Overinvestment in AI by China and underinvestment by the U.S. could lead to China’s superior capabilities. The strategic importance of Taiwan’s TSMC and China’s potential military ambitions highlight AI’s critical role in global power dynamics.

AI’s Potential to Surpass Human Control:

Some AI leaders predict that AI algorithms could outthink humans in years, raising concerns about AI’s controllability. Alexandr Wang considers this fear unreasonable based on current technology and understanding. Experts raise concerns about AI surpassing human intelligence, though current technology suggests this fear is premature.

AI in Military Power Dynamics:

AI is poised to revolutionize military power, with the Chinese Communist Party (CCP) keenly investing in AI to challenge U.S. dominance. The U.S. DoD has established ethical AI principles to guide the development and use of AI systems. Alexandr Wang believes technologies should adhere to these principles. Testing and evaluation of AI systems are crucial to ensure maturity and minimize risks on the battlefield. Some autonomous weapon systems with limited decision-making capabilities are already in use. However, humans are typically still in the loop, and the technology serves as a decision aid rather than a decision-maker. The challenge lies in synthesizing overwhelming information from diverse sources into actionable intelligence for military leaders. AI’s role as a decision aid, backed by extensive human involvement in data labeling and algorithm development, is crucial.

Ethical Considerations and Principles for AI Deployment:

The U.S. DoD’s ethical AI principles guide responsible AI use, emphasizing the need to avert risks associated with premature AI deployment. Ensuring AI system maturity through rigorous testing is crucial for effective and ethical deployment in sensitive areas like military use. The current state of autonomous weapon systems indicates that they are capable of decision-making, but human oversight remains central to the decision-making process in military contexts.

Global Proliferation and Concerns of AI:

The swift spread of AI technologies worldwide, including China’s advancements, raises concerns about potential misuse and the need for responsible governance. Global proliferation of AI technology is a concerning trend. Open-source models are being developed in various countries, including France, the UAE, and China. Humans will always play a critical role in AI development. Humans will be involved in developing algorithms and providing input and preferences to guide model behavior. Reinforcement learning with human feedback (RHF) has improved the helpfulness of AI models and ensured they adhere to human intent and principles. The military faces challenges in synthesizing information from various sensors and platforms into actionable intelligence. Autonomous weaponry and operations are not the focus; the emphasis is on decision aid and helping human decision-makers and operators.



In conclusion, Scale’s trajectory in AI infrastructure and its strategic positioning amidst geopolitical tensions underscore the multifaceted challenges and opportunities in the AI landscape. While AI promises revolutionary changes in various sectors, especially military, the ethical, practical, and geopolitical implications of its rapid development and deployment demand cautious, well-informed approaches. The collaborative efforts of governments, corporations, and ethical bodies are crucial in navigating the complex tapestry of AI’s global impact.

Notes by: BraveBaryon