DW Weekly #200 – AI’s new Cold War redraws geopolitical lines
The evolution of AI, particularly the race to develop cutting-edge large language models (LLMs), has become an arena for geopolitical manoeuvring, intersecting with data privacy, technological sovereignty, and international relations. Nations, including the USA, China, South Korea, and the EU, are investing heavily in AI technology, seeking dominance in a field that intersects deeply with global power dynamics, data privacy, and national security concerns.
At the centre of this struggle in the last two months is China’s DeepSeek, a company that has adopted a disruptive open-source philosophy. Unlike Western proprietary models such as OpenAI’s ChatGPT and xAI’s Grok-3, DeepSeek provides unrestricted access to its AI architecture, allowing developers worldwide to modify and self-host its system. This setup fosters transparency, reduces dependence on expensive computing power, and raises critical security concerns. The problem is that Chinese data protection laws could compel data sharing with government entities, sparking fears over data sovereignty and international privacy breaches. These concerns have led countries like South Korea to block new downloads of DeepSeek, signalling the growing unease around AI systems developed under Chinese regulatory frameworks.
South Korea’s ban on DeepSeek exemplifies how AI development is entangled with national security. A global leader in semiconductors and AI research, South Korea has worked closely with US firms in AI advancement, aligning itself with Washington’s broader efforts to curb Chinese technological expansion. Meanwhile, North Korea remains a wildcard in this AI race, having already demonstrated its cyberwarfare capabilities. South Korea’s proactive stance on AI reflects not only its economic ambitions but also its strategic imperative to counter threats from its northern neighbour and the wider geopolitical tensions stemming from China’s AI rise.
DeepSeek’s open-source approach also sets it apart in the competition between transparency and control. Western AI leaders, including OpenAI and xAI, operate in a more restricted environment where proprietary systems are developed under stringent regulatory oversight. OpenAI, for example, enforces strict data security measures and maintains tight control over its AI outputs, ensuring compliance with international laws. In contrast, DeepSeek’s more accessible framework democratises AI development but leaves room for potential misuse, from security vulnerabilities to circumventing global content moderation policies.
In contrast to DeepSeek’s open model, Grok-3, led by Elon Musk’s xAI, represents a closed, exclusive approach that prioritises performance and real-time adaptability. Grok-3’s DeepSearch feature enhances its reasoning capabilities, allowing users to see how the AI formulates responses. While Musk touts Grok-3 as a breakthrough in AI evolution, its limited accessibility—restricted to X’s Premium+ subscribers and a forthcoming SuperGrok subscription—positions it within an elite, controlled market rather than an open AI ecosystem like DeepSeek’s.
The broader question at the heart of the AI arms race is data governance. The EU’s AI Act, the USA’s evolving regulations, and China’s cybersecurity laws all attempt to strike a balance between fostering innovation and maintaining control. Countries are increasingly aware that AI is more than just a technological tool—it is an instrument of geopolitical influence, capable of shifting economic power, enhancing intelligence operations, and altering global ideological narratives. The fact that AI models can be weaponised to manipulate information or infringe upon national security makes their regulation an urgent international priority.
As AI technology advances at breakneck speed, the fundamental debate between open-source accessibility and proprietary control will shape the global AI landscape. Whether China’s DeepSeek succeeds in reshaping AI innovation through transparency or whether proprietary models like OpenAI and xAI maintain dominance through closed systems, the future of AI remains deeply entangled with the world’s geopolitical chessboard. The race for LLM supremacy is not just about technological breakthroughs—it is about control, security, and the foundation of digital power in the 21st century.
The data protection issue is not confined to South Korea; Italy previously blocked DeepSeek’s chatbot due to privacy policy concerns.
Trump discusses TikTok sale with China
President Donald Trump confirmed on Wednesday that he was in active discussions with China over the future of TikTok, as the US seeks to broker a sale of the popular app.
Australian kids overlook social media age checks
A new report reveals that the country faces challenges enforcing social media age restrictions for children amidst upcoming bans.
Marko and the Digital Watch team
Highlights from the week of 14-21 February 2025
This year’s summit showcased cutting-edge innovations, notably featuring Physical AI robots and Boston Dynamics’ Spot robot dogs.
In a significant step towards AI integration, The New York Times has granted its editorial and product teams permission to use artificial intelligence tools.
Critics claim Russia is deliberately slowing YouTube to limit access to critical content.
The AI industry is witnessing substantial financial investment, with xAI and OpenAI pursuing significant fundraising efforts. Emerging competitors, such as China’s DeepSeek, with cost-efficient models, indicate potential disruptions in the…
Harvard physicist hails Microsoft’s quantum breakthrough as a potential leap forward.
A federal court has ruled that Elon Musk’s Department of Government Efficiency (DOGE) can continue accessing US government agency data for now, despite concerns over its authority.
Disinformation grows as young Europeans turn their back to traditional media.
Inspired by DeepSeek’s AI success, Baidu intends to release Ernie 4.5 as an open-source model.
Financial fundraising efforts, highlighting the company’s pursuit to uphold its purchase valuation.
CHIPS Act terms face reconsideration under the Trump administration, potentially causing delays in semiconductor subsidies.
A diplomatic row has erupted between Mexico and Google over the renaming of the Gulf of Mexico on US versions of Google Maps.
www.diplomacy.edu
Ethics provides a framework for informed decision-making and distinguishing right from wrong. While technology is neutral, ethical dilemmas arise from human design and use of AI systems. Is “AI ethics” misleading since individuals hold ultimate responsibility for ethical behaviour?
www.diplomacy.edu
Did Indian independence leader Subhas Chandra Bose get an unfair reputation? From challenging caste divisions to empowering women, his vision was more radical than many realise. Aldo Matteucci explores.
www.diplomacy.edu
The CitiVerse Initiative is testing how AI, digital twins, and virtual worlds can enhance urban life. Can this technology-driven approach create more inclusive and sustainable cities? Dr Anita Lamprecht explores
www.diplomacy.edu
Young Serbians are transforming nationalism from division to unity through digital activism and inclusive ideals. Learn about Nation 4.0, a movement redefining Serbian identity.
www.diplomacy.edu
In today’s digital world, protecting state data is essential. Data embassies, supported by the Vienna Convention, provide an innovative solution. Countries like Estonia and Monaco use them to enhance cybersecurity, ensuring data remains safe and secure.
Tech attache briefing: Policy implications and international landscape The event is part of a series of regular briefings on the Geneva Internet Platform (GIP)
