5 Tech Perspectives to Ponder

Insights from US-China relations seminar in Hong Kong

This week marks the start of China’s annual parliamentary meetings, known as lianghui (两会) or the Two Sessions in Beijing. In a similar way that November election season holds significance to Americans, lianghui is a moment of economic and political consequence in China. Last week, I attended a conference on US-China relations hosted by the National Committee of US-China Relations and the Schwarzman Scholars Program. While the speakers were participating under Chatham House Rules, the collective experience of attending gave rise to some interesting takeaways. So instead of a long-form deep dive into a single topic, I wanted to share 5 perspectives to ponder from my reflections on US-China last week:

1. Chinese tech regulations are much more lax than they seem.

2. US “crackdown” policies are gifts in disguise.

3. We’re on the cusp of a hardware renaissance.

4. Data sovereignty matters now more than ever.

5. Geopolitical competition can be jet fuel for progress.

👮‍♀️ Chinese tech regulations are more lax than they seem.

People often view China as a tough regulator - likely because of the influence of the Chinese Communist Party. But this is a far cry from reality. What China cares about is content control. The government is fundamentally insecure about public discourse and concerned that the output of generative AI could be politically misaligned with its socialist agenda. Apart from this, China’s national agenda is to encourage the growth and development of its tech sector (like pretty much everyone else).

The AI regulations from the Cyberspace Administration of China (CAC) are a case in point. CAC made waves when they first announced Draft Measures to regulate AI. Early measures implemented a mandatory licensing scheme and included provisions that would require AI companies to “uphold core social values” in their products (see Article 4, Section 1). Later in 2023, the CAC released Interim Measures which substantially amended the Draft Measures and made them much more accommodating and business friendly. Later versions removed requirements to ensure truth and accuracy of information, verify user’s real identity, and prohibit discrimination. Today, China’s law only applies to public-facing generative AI systems. In essence, what began as a formal licensing model has evolved into a watered-down registration system.

Already, this AI leniency has its costs. Last month, an unidentified Hong Kong multinational lost a staggering $25.6 million in a deepfake scam. At the Hong Kong office, an employee was instructed to transfer funds by a digitally recreated version of the company CFO along with several other fake participants. The scammers turned publicly available video footage into convincing versions of all meeting participants so everyone on the video call was fake except for the victim. Stories like this underscore how AI trust and safety is a deeply unsolved global problem. In a world where Chinese domestic regulation is weak, international governance is needed as a complement/substitute to nudge China and other global powers to do more.

🎁 US “crackdown” policies are gifts in disguise.

The US has tried to respond. But American AI regulation is widely regarded as performative. Biden’s AI Executive Order does little besides urge for the development of trustworthy AI. Other US policies are not only symbolic, but have also been counterproductive. Placing Chinese companies on sanctions lists and subjecting them to heightened export controls are often framed as crackdowns to deter China. But in practice, these policies are gifts in disguise.

In 2022, Biden signed the CHIPS and Science Act which provides $50+ billion to spur American semiconductor research, development, manufacturing, and workforce development. The majority or roughly $39 billion is earmarked for manufacturing incentives. $13.2 billion are intended for R&D and workforce development, $2 billion are intended for legacy chips used in automobiles and defense systems, and $500 million are for international information communications technology security and semiconductor supply chain activities. An explicit goal of CHIPS is to “Counter China”. However, somewhat counterintuitively, this and recent US chip restrictions have supported the growth of Chinese companies. It is forcing China to learn to be more self sufficient, and that could be bad for long-run US interests.

🛠️ We’re on the cusp of a hardware renaissance.

The last century of growth and development was driven by software. See “software is eating the world”. Increasingly, though, it seems like opportunities for software-based improvements are becoming scarce. The truly disruptive development opportunity for the next generation is hardware. Here, China has a substantial leg up. Places like Shenzhen are already manufacturing heartlands. The city has a vibrant community of hardware-oriented hackers, maker spaces, incubators, and accelerators. In many ways, the US is playing catch up. A look at recent AI headlines serves as a clear indicator: CHIPS is a concerted effort to boost US manufacturing capacity, NVIDIA stock is soaring, and companies like SambaNova and Groq have garnered attention for hardware/chipmaking advances. The tech community is realizing that the current generation of AI is bottlenecked by silicon, not code.

Beyond chips, hardware is resurgent across the broader tech ecosystem: cars, batteries, wearables like the Vision Pros, and even art. Physical technologies are what will uplift global economies into recovery. Innovation is no longer about the digital economy alone. In time, Silicon Valley and Europe’s software hubs will align with Asias manufacturing industry. And the software that changes and enhances physical reality seems well positioned.

👨‍💻 Data sovereignty matters a lot more now.

Data sovereignty addresses the ownership, control and legal aspects of data. The central question is: who owns personal data? Today, the law provides no clear answer. Instead, most countries/states have enacted privacy laws to delineate the boundaries of what’s permissible. The California Consumer Privacy Act of 2018 is one of the most ambitious data privacy legislation ever enacted in the US. The EU has the General Data Protection Regulation (GDPR). China has the Personal Information Protection Law (PIPL). You don’t own your data, but you can protect it.

Some have argued that in order to have more inclusive growth, the economic value of individual data needs to be recognized. China seems well on its way. In 2020, China’s State Council announced that data is one of its five factors of production. The other four were capital, land, talent, and technology. China’s 14th Five Year Plan aims to “activate the factor value of data” by creating a national open data platform for public data and encouraging private businesses to share proprietary data. Since data is produced by every individual simply by virtue of their interaction with the economic system, monetizing it at scale creates a unique opportunity for economic uplift that is accessible to all. This universality is especially appealing to China, which has the second largest total GDP in the world but is way behind in per capita terms. Importantly, the American middle class has been shrinking, while China’s middle class is growing. As a result, China is likely more open to ideas about how to create economic uplift by helping individuals own and monetize their personal data. China is one of the most digitalized countries in the world so its per capita income could leapfrog ahead if individual data sovereignty was implemented.

🚀 Geopolitical competition can be jet fuel for progress.

We’re entering an era of strong competition and tense relations between the US and China. Many have predicted this will last at least a decade. But what if that was viewed as a gift rather than a risk? When we think of America's Cold War competition with the Soviet Union, we often remember a time of intense hostility and nuclear armament. But this competition also set the stage for the Great Space Race - a fast-paced era of scientific productivity that we still benefit from today. Today, as we stand at the cusp of a similarly competitive period with a rising China, it’s worth considering how we might guide competition towards productive ends. As rapid development of AI places unprecedented strain on our energy infrastructure, might this finally be an occasion for major superpowers to invest heavily in new and sustainable energy sources? Could the needs of AI and global one-upmanship serve as the economic and scientific catalyst for an energy revolution? While there’s plenty to be concerned about when considering global competition, let’s not forget the substantial potential for positive externalities as well.