ARTIFICIAL INTELLIGENCE

“I do think countries all should invest in their own human capital, invest in partnerships and invest in their own technological stack as well as the business ecosystem… I think not investing in AI would be macroscopically the wrong thing to do.” – Dr. Fei-Fei Li – Stanford Professor – world-renowned authority in artificial intelligence

The statement was delivered during a high-stakes panel discussion on artificial superintelligence, convened at the Future Investment Initiative in Riyadh, where nation-state leaders, technologists, and investors gathered to assess their strategic positioning in the emerging AI era. Her words strike at the heart of a dilemma facing governments worldwide: how to build national AI capabilities whilst avoiding the trap of isolationism, and why inaction would be economically and strategically untenable.

Context: The Geopolitical Stakes of AI Investment

The Historical Moment

Dr. Li’s statement comes at a critical juncture. By late 2024 and into 2025, artificial intelligence had transitioned from speculative technology to demonstrable economic driver. Estimates suggested AI could generate between $15 trillion and $20 trillion in economic value globally by 2030—a figure larger than the current gross domestic product of most nations. This windfall is not distributed evenly; rather, it concentrates among early movers with capital, infrastructure, and talent. The race is on, and the stakes are existential for national competitiveness, employment, and geopolitical influence.

In this landscape, a nation that fails to invest in AI capabilities risks profound economic displacement. Yet Dr. Li is equally clear: isolation is counterproductive. The most realistic path forward combines three pillars:

• Human Capital: The talent to conceive, build, and deploy AI systems
• Partnerships: Strategic alliances, particularly with leading technological ecosystems (the US hyperscalers, for instance)
• Domestic Technological Infrastructure: The local research bases, venture capital, regulatory frameworks, and business ecosystems that enable sustained innovation

This is not a counsel of surrender to Silicon Valley hegemony, but rather a sophisticated argument about comparative advantage and integration within global technological networks.

Dr. Fei-Fei Li: The Person and Her Arc

Early Life and Foundational Values

Dr. Fei-Fei Li’s perspective is shaped by her personal trajectory. Born in Chengdu, China, she emigrated to the United States at age fifteen, settling in New Jersey where her parents ran a small business. This background infuses her thinking: she understands both the promise of technological mobility and the structural barriers that constrain developing economies. She obtained her undergraduate degree in physics from Princeton University in 1999, with high honours, before pursuing doctoral studies at the California Institute of Technology, where she worked across computer science, electrical engineering, and cognitive neuroscience, earning her PhD in 2005.

The ImageNet Revolution

In 2007, whilst at Princeton, Dr. Li embarked on a project that would reshape artificial intelligence. Observing that cognitive psychologist Irving Biederman estimated humans recognise approximately 30,000 object categories, Li conceived ImageNet: a massive, hierarchically organised visual database. Colleagues dismissed the scale as impractical. Undeterred, she led a team (including Princeton professors Jia Deng, Kai Li, and Wei Dong) that leveraged Amazon Mechanical Turk to label over 14 million images across 22,000 categories.

By 2009, ImageNet was published. More critically, the team created the ImageNet Large Scale Visual Recognition Challenge (ILSVRC), an annual competition that invited researchers worldwide to develop algorithms for image classification. This contest became the crucible in which modern deep learning was forged. When Geoffrey Hinton’s group achieved a breakthrough using convolutional neural networks in 2012, winning the competition by a decisive margin, the deep learning revolution was catalysed. ImageNet is now recognised as one of the three foundational forces in the birth of modern AI.

What is instructive here is that Dr. Li’s contribution was not merely technical but infrastructural: she created a shared resource that democratised AI research globally. Academic groups from universities across continents—not just Silicon Valley—could compete on equal footing. This sensibility—that progress depends on enabling distributed talent—runs through her subsequent work.

Career Architecture and Strategic Leadership

Following her Princeton years, Dr. Li joined Stanford University in 2009, eventually becoming the Sequoia Capital Professor of Computer Science—a title of singular prestige. From 2013 to 2018, she directed Stanford’s Artificial Intelligence Lab (SAIL), one of the world’s premier research institutes. Her publications exceed 400 papers in top-tier venues, and she remains one of the most cited computer scientists of her generation.

During a sabbatical from Stanford (January 2017 to September 2018), Dr. Li served as Vice President and Chief Scientist of AI/ML at Google Cloud. Her mandate was to democratise AI technology, lowering barriers for businesses and developers—work that included advancing products like AutoML, which enabled organisations without deep AI expertise to deploy machine learning systems.

Upon returning to Stanford in 2019, she became the founding co-director of the Stanford Institute for Human-Centered Artificial Intelligence (HAI), an explicitly multidisciplinary initiative spanning computer science, social sciences, humanities, law, and medicine—all united by the conviction that AI must serve human flourishing, not vice versa.

Current Work and World Labs

Most recently, Dr. Li co-founded and serves as chief executive officer of World Labs, an AI company focused on spatial intelligence and generative world models. This venture extends her intellectual agenda: if large language models learn patterns over text, world models learn patterns over 3D environments, enabling machines to understand, simulate, and reason about physical and virtual spaces. For robotics, healthcare simulation, autonomous systems, and countless other domains, this represents the next frontier.

Recognition and Influence

Her standing is reflected in numerous accolades: election to the National Academy of Engineering, the National Academy of Medicine (2020), and the American Academy of Arts and Sciences (2021); the Intel Lifetime Achievement Innovation Award in 2023; and inclusion in Time magazine’s 100 Most Influential People in AI. She is colloquially known as the “Godmother of AI.” In 2023, she published a memoir, The Worlds I See: Curiosity, Exploration and Discovery at the Dawn of AI, which chronicles her personal journey and intellectual evolution.

Leading Theorists and Strategic Thinkers: The Landscape of AI and National Strategy

The backdrop to Dr. Li’s statement includes several strands of thought about technology, development, and national strategy:

Economic and Technological Diffusion

• Erik Brynjolfsson and Andrew McAfee (The Second Machine Age, Machine Platform Crowd): These MIT researchers have articulated how technological revolutions create winners and losers, and that policy choices—not technology alone—determine whether gains are broadly shared. They underscore that without intentional intervention, automation and AI tend to concentrate wealth and opportunity.
• Dani Rodrik (Harvard economist): Rodrik’s work on “premature demonetarisation” and structural transformation highlights the risks faced by developing economies when technological progress accelerates faster than institutions can adapt. His analysis supports Dr. Li’s argument: countries must actively build capacity or risk being left behind.
• Mariana Mazzucato (University College London): Mazzucato’s research on the entrepreneurial state emphasises that breakthrough innovations—including AI—depend on public investment in foundational research, education, and infrastructure. Her work buttresses the case for public and private sector partnership.

Artificial Intelligence and Cognition

• Geoffrey Hinton, Yann LeCun, and Yoshua Bengio: The triumvirate of deep learning pioneers recognised that neural networks could scale to superhuman performance in perception and pattern recognition, yet have increasingly stressed that current approaches may be insufficient for general intelligence. Their candour about limitations supports a measured, long-term investment view.
• Stuart Russell (UC Berkeley): Russell has been a prominent voice calling for AI safety and governance frameworks to accompany capability development. His framing aligns with Dr. Li’s insistence that human-centred values must guide AI research and deployment.

Geopolitics and Technology Competition

• Michael Mazarr (RAND Corporation): Mazarr and colleagues have analysed great-power competition in emerging technologies, emphasising that diffusion of capability is inevitable but the pace and terms of diffusion are contestable. Nations that invest in talent pipelines and partnerships will sustain influence; those that isolate will atrophy.
• Kai-Fu Lee: The Taiwanese-American venture capitalist and author (AI Superpowers) has articulated how the US and China are in a competitive race, but also how smaller nations and regions can position themselves through strategic partnerships and focus on applied AI problems relevant to their economies.
• Eric Schmidt (former Google CEO): Schmidt, who participated in the same FII panel as Dr. Li, has emphasised that geopolitical advantage flows to nations with capital markets, advanced chip fabrication (such as Taiwan’s TSMC), and deep talent pools. Yet he has also highlighted pathways for other nations to benefit through partnerships and focused investment in particular domains.

Human-Centred Technology and Inclusive Growth

• Timnit Gebru and Joy Buolamwini: These AI ethics researchers have exposed how AI systems can perpetuate bias and harm marginalised communities. Their work reinforces Dr. Li’s emphasis on human-centred design and inclusive governance. For developing nations, this implies that AI investment must account for local contexts, values, and risks of exclusion.
• Turing Award recipients and foundational figures (such as Barbara Liskov on systems reliability, and Leslie Valiant on learning theory): Their sustained emphasis on rigour, safety, and verifiability underpins the argument that sustainable AI development requires not just speed but also deep technical foundations—something that human capital investment cultivates.

Development Economics and Technology Transfer

• Paul Romer (Nobel laureate): Romer’s work on endogenous growth emphasises that ideas and innovation are the drivers of long-term prosperity. For developing nations, this implies that investment in research capacity, education, and institutional learning—not merely adopting foreign technologies—is essential.
• Ha-Joon Chang: The heterodox development economist has critiqued narratives of “leapfrogging” technology. His argument suggests that nations building indigenous technological ecosystems—through domestic investment in research, venture capital, and entrepreneurship—are more resilient and capable of adapting innovations to local needs.

The Three Pillars: An Unpacking

Dr. Li’s framework is sophisticated precisely because it avoids two traps: technological nationalism (the fantasy that any nation can independently build world-leading AI from scratch) and technological fatalism (the resignation that small and medium-sized economies cannot compete).

Human Capital

The most portable, scalable asset a nation can develop is talent. This encompasses:

• Education pipelines: From primary through tertiary education, with emphasis on mathematics, computer science, and critical thinking
• Research institutions: Universities, national laboratories, and research councils capable of contributing to fundamental and applied AI knowledge
• Retention and diaspora engagement: Policies to keep talented individuals from emigrating, and mechanisms to attract expatriate expertise
• Diversity and inclusion: As Dr. Li has emphasised through her co-founding of AI4ALL (a nonprofit working to increase diversity in AI), innovation benefits from diverse perspectives and draws from broader talent pools

Partnerships

Rather than isolating, Dr. Li advocates for strategic alignment:

• North-South partnerships: Developed nations’ hyperscalers and technology firms partnering with developing economies to establish data centres, training programmes, and applied research initiatives. Saudi Arabia and the UAE have pursued this model successfully
• South-South cooperation: Peer learning and knowledge exchange among developing nations facing similar challenges
• Academic and research collaborations: Open-source tools, shared benchmarks (as exemplified by ImageNet), and collaborative research that diffuse capability globally
• Technology licensing and transfer agreements: Mechanisms by which developing nations can access cutting-edge tools and methods at affordable terms

Technological Stack and Ecosystem

A nation cannot simply purchase AI capability; it must develop home-grown institutional and commercial ecosystems:

• Open-source communities: Participation in and contribution to open-source AI frameworks (PyTorch, TensorFlow, Hugging Face) builds local expertise and reduces dependency on proprietary systems
• Venture capital and startup ecosystems: Policies fostering entrepreneurship in AI applications suited to local economies (agriculture, healthcare, manufacturing)
• Regulatory frameworks: Balanced approaches to data governance, privacy, and AI safety that neither stifle innovation nor endanger citizens
• Domain-specific applied AI: Rather than competing globally in large language models, nations can focus on AI applications addressing pressing local challenges: medical diagnostics, precision agriculture, supply-chain optimisation, or financial inclusion

Why Inaction Is “Macroscopically the Wrong Thing”

Dr. Li’s assertion that not investing in AI would be fundamentally mistaken rests on several converging arguments:

Economic Imperatives

AI is reshaping productivity across sectors. Nations that fail to develop internal expertise will find themselves dependent on foreign technology, unable to adapt solutions to local contexts, and vulnerable to supply disruptions or geopolitical pressure. The competitive advantage flows to early movers and sustained investors.

Employment and Social Cohesion

While AI will displace some jobs, it will create others—particularly for workers skilled in AI-adjacent fields. Nations that invest in reskilling and education can harness these transitions productively. Those that do not risk deepening inequality and social fracture.

Sovereignty and Resilience

Over-reliance on foreign AI systems limits national agency. Whether in healthcare, defence, finance, or public administration, critical systems should rest partly on domestic expertise and infrastructure to ensure resilience and alignment with national values.

Participation in Global Governance

As AI governance frameworks emerge—whether through the UN, regional bodies, or multilateral forums—nations with substantive technical expertise and domestic stakes will shape the rules. Those without will have rules imposed upon them.

The Tension and Its Resolution

Implicit in Dr. Li’s statement is a tension worth articulating: the world cannot support 200 competing AI superpowers, each building independent foundational models. Capital and talent are finite. Yet neither is the world a binary of a few AI leaders and many followers. The resolution lies in specialisation and integration:

• A nation may not lead in large language models but excel in robotics for agriculture
• It may not build chips but pioneer AI applications in healthcare or education
• It may not host hyperscaler data centres but contribute essential research in AI safety or fairness
• It will necessarily depend on global partnerships whilst developing sovereign capacity in domains critical to its citizens

This is neither capitulation nor isolation, but rather a mature acceptance of global interdependence coupled with strategic autonomy in domains of national importance.

Conclusion: The Compass for National Strategy

Dr. Li’s counsel, grounded in decades of research leadership, industrial experience, and global perspective, offers a compass for policymakers navigating the AI era. Investment in human capital, strategic partnerships, and home-grown technological ecosystems is not a luxury or academic exercise—it is fundamental to national competitiveness, prosperity, and agency. The alternative—treating AI as an external force to be passively absorbed—is indeed “macroscopically” mistaken, foreclosing decades of economic opportunity and surrendering the right to shape how this powerful technology serves human flourishing.