THIS WEEK IN AI GOVERNANCE

The latest in defense, regulation, gov tech & geopolitics

• AI’s limitations may shape geopolitics more than its capabilities

• India’s AI regulation for securities markets falls short, putting retail investors at risk

• EU commits €1.3 billion to boost digital sovereignty through 2027

• Leaked database reveals China’s AI-powered censorship system targeting political content

• Enchantment Under the Sea: Meta’s 50,000-km cable project to boost connectivity across continents

• Wooing with ooze: UAE uses oil wealth to secure strategic AI partnerships with the US

• NY court rejects AI avatar in courtroom as judges crack down on digital deception

• Washington school district transforms AI policy with teacher-guided framework for education

• Lockheed Martin partners with Google Cloud to bring generative AI to defense operations

• North Korea unveils AI-equipped suicide drones amid deepening Russia ties

Funding AI education, industry immersion, and career assistance to those who have served our nation.

What’s happening in AI policy right now

China cracks down as Europe invests

_{New technologies bring both opportunities and challenges as global players advance their AI strategies}

This week brings significant developments in the AI landscape, with new initiatives that illustrate how different regions are approaching artificial intelligence development and regulation.

A leaked database has revealed China's development of a sophisticated AI-powered censorship system that uses large language model technology to detect politically sensitive content with unprecedented accuracy. This system represents a significant evolution in digital control, moving beyond simple keyword filtering to identify subtle expressions of dissent—even those craftily embedded in traditional Chinese idioms implying regime instability.

In separate news, the European Commission has announced a €1.3 billion investment package through its Digital Europe Programme for 2025-2027, focusing on strategic digital technologies to enhance Europe's technological sovereignty and competitiveness. This funding will prioritize AI deployment, cybersecurity, and digital skills development.

_{Advanced AI-powered censorship in China}

China's new censorship system is trained on over 133,000 examples of "sensitive content" and flags material by priority level, with military affairs, Taiwan-related discussions, and political criticism receiving the highest priority. Evidence of implementation can already be seen in Chinese AI platforms like the DeepSeek chatbot, which demonstrates active political censorship.

The implications extend far beyond mere content moderation—DeepSeek's release as a foundation model raises concerns about AI systems being weaponized for propaganda purposes. As a base for training other AI systems, such models could potentially spread state-approved worldviews across borders, creating a new frontier in global information warfare.

This development represents what OpenAI CEO Sam Altman has described as an ideological dimension in AI development. China's censorship system demonstrates a practical application where AI's pattern recognition capabilities are being harnessed for content control—showing how AI can be extremely capable in specific domains while serving particular strategic objectives.

As Ethan Mollick notes in his analysis of AI discussions, we often fall into a "weird dichotomy" where AI is either dismissed as mere hype or feared as the imminent arrival of superhuman machines. But the reality lies in the nuanced middle ground where AI proves extremely capable in some domains while remaining limited in others.

_{Europe's strategic digital investment}

The European Union is making substantial investments in its digital future. The €1.3 billion investment approved for 2025-2027 prioritizes AI deployment, cybersecurity, and digital skills development, with funding opportunities becoming available to various European stakeholders starting April 2025.

This approach aligns with Clayton Christensen's concept of "jobs to be done"—the EU has identified specific problems it needs to solve, including bolstering cybersecurity infrastructure, addressing digital skills gaps, and deploying AI in ways that maintain European values and competitiveness. Rather than viewing AI as either a panacea or a threat, Europe is focused on the practical applications that serve its strategic needs.

The EU's investment comes amid growing global concerns about AI regulation and sovereignty. While technological development continues at breakneck speed, the political and regulatory frameworks are struggling to keep pace. Europe's approach attempts to resolve this tension by investing in both technology and the human capital needed to deploy it effectively.

_{Building competitive moats in AI}

These developments highlight what venture capitalist Jerry Chen calls "the new moats" in the AI landscape. Traditional competitive advantages are being reshaped by how organizations develop, deploy, and regulate AI systems.

In China's case, the development of sophisticated AI censorship systems represents a form of technological moat—creating capabilities that align with specific strategic objectives. The training of large language models on carefully curated datasets creates systems that behave in ways that serve particular national interests.

For Europe, the investment in digital sovereignty represents a different kind of moat—one based on creating an environment where innovation can flourish while maintaining alignment with European values and priorities. This approach focuses on building both technological capabilities and human capital.

As Marc Andreessen argues in his provocative essay "Why AI Will Save the World," AI offers tremendous potential to solve human problems; however, how AI is deployed will vary based on the values and priorities of the societies creating it.

_{The emerging AI landscape}

The global AI landscape is evolving rapidly, with major implications for businesses, governments, and society. The technological capabilities are advancing faster than our understanding of their implications or our ability to create appropriate governance frameworks.

As Sam Harris warns in his discussion of potential "digital apocalypse" scenarios, the values embedded in AI systems will ultimately determine whether they benefit humanity broadly or serve narrower interests. The capabilities being developed today—from content moderation to data analysis—will shape the digital environment for years to come.

For businesses and organizations navigating this complex environment, understanding regional approaches to AI becomes critical. Companies operating globally must consider how their AI strategies align with local regulations and values, potentially requiring different approaches in different markets.

_{Looking ahead}

As AI technologies continue to develop, we can expect to see increasing specialization and sophistication in how they're deployed. The examples of China's censorship systems and Europe's strategic investments represent just the beginning of what will likely be increasingly diverse applications.

These developments raise important questions for the future:

• How will AI systems trained on different value systems and for different purposes interact?

• What governance frameworks will emerge to address cross-border AI applications?

• How will businesses navigate an increasingly complex global AI landscape?

• What new capabilities will emerge as AI technologies continue to advance?

The answers to these questions will shape not just the future of AI technology but potentially the structure of global information flows and economic activity for decades to come.

As Max Tegmark reminds us in his cautionary writing on AI risks, the decisions we make now about AI development and governance will have profound implications for humanity's future. The approaches different regions are taking represent not just different regulatory frameworks, but fundamentally different visions of technology's role in society.