DeepSeek's low-cost AI fuels debates on democratisation but raises concerns over China's pursuit of strategic dominance
The emergence of the DeepSeek R1 model has reignited debates around the democratisation of artificial intelligence (AI), a movement focused on advancing the AI ecosystem by making AI technologies more accessible, accountable, affordable, transparent, and inclusive.
This development raises critical questions about competition, regulations, ethics, and transparency in AI's future. The emergence of DeepSeek has intensified debates around AI democratisation more than ChatGPT, despite ChatGPT's earlier and widespread adoption, mainly due to DeepSeek's disruptive cost efficiency and open-source approach. While OpenAI spent hundreds of millions developing ChatGPT, DeepSeek's R1 model was built for just US$5-6 million, demonstrating that high-performance AI can be achieved at a fraction of the cost. This challenges Silicon Valley's technological monopoly.
DeepSeek provided a fully open-source model relative to ChatGPT's restrictive licensing and undisclosed training data. Moreover, DeepSeek's success has fuelled discussions on technological sovereignty, presenting the Chinese model of state-supported, self-reliant AI development as a viable alternative to the US-controlled AI ecosystem.
The emergence of DeepSeek has intensified debates around AI democratisation more than ChatGPT, despite ChatGPT's earlier and widespread adoption, mainly due to DeepSeek's disruptive cost efficiency and open-source approach.
Washington-based global investment management firm Bernstein's research estimates that "DeepSeek's pricing is 20 to 40 times cheaper than equivalent models," a cost advantage that could be a game-changer for startups, smaller firms, and independent developers. Beyond affordability, DeepSeek emphasises user control and self-hosting options, aligning its approach with democratic principles.
The focus on empowering and democratising AI, along with the affordability of AI technologies, could have long-term implications for reducing AI biases, enhancing transparency and strengthening robust governance frameworks, which is key towards democratisation of the AI ecosystem.
The democratisation of AI is often justified by its potential to enhance competitiveness, drive economic growth, and advance problem-solving capabilities across critical domains, from healthcare and education to governance, science, and engineering.
The democratisation of AI carries significant risk for state legitimacy, as democratic government fundamentally depends on it. While democracy serves as a common ground in pluralistic societies, AI governance often neglects reasonable pluralism due to resource constraints. Existing regulatory frameworks spanning criminal justice, healthcare, consumer protection, and education must adapt to AI's role in these domains.
However, AI deployment at the government level raises deeper questions about the broader reform, creating a risk of a parallel system of authority by the governance unless carefully integrated into decentralised governance structures. At worst, it may exacerbate inequality due to structural barriers (e.g., lobbying advantage, corporate resources) disproportionately excluding low-income and minorities, undermining inclusivity. AI-based crowdsource participation, often endorsed as a democratic solution, tends to simplify the myopic and shallow input, sidelining the marginalised voices and long-term concerns like climate security, future generations, and non-citizens (e.g., migrants and the global poor). The rush for democratisation of AI frequently prioritises crowdsourcing participation over deliberative quality, favouring short-term, parochial interest over systemic justice.
AI deployment at the government level raises deeper questions about the broader reform, creating a risk of a parallel system of authority by the governance unless carefully integrated into decentralised governance structures.
The US-China strategic rivalry, particularly in AI innovation & development, further complicates the global technological and geopolitical dynamics. DeepSeek, a Chinese AI chatbot, has drawn comparisons with existing models like ChatGPT, raising critical questions about its role in the narrative of AI democratisation.
China's coercive actions and ambition to become a "cyber superpower" contrast sharply with its domestic policies of mass surveillance, censorship, and the Great Firewall. This duality fuels scepticism about whether China's promotion of AI abroad aligns with democratic values or serves a false narrative of democratisation to expand its influence. Beyond democratisation, China's true objectives seem to focus on shaping international norms and achieving dominance through intelligence, surveillance, and reconnaissance, heightening the tensions with the US.
A key concern surrounding DeepSeek is data privacy risks, given its open-access nature. These concerns have led to bans by Taiwan, Texas, NASA, and the US Navy, with ongoing investigations in Greece, Belgium, Ireland, France, and Australia. Additionally, countries like Australia, South Korea, and India have restricted DeepSeek's use on government devices, reflecting growing apprehension over potential security threats posed by Chinese AI technologies.
Additionally, technology functions as a strategic tool for shaping and controlling adversarial behaviour through soft power. Operating cost-effectively within China's restricted information ecosystem, it suppresses discussion on sensitive and anti-China topics, raising concerns about AI-driven bias and propaganda dissemination.
Democratic values, however, extend beyond mere freedom of speech; they demand neutrality and unbiased information. While Chinese state-led companies, such as DeepSeek, claim to amplify effectively marginalised voices, this assertion is misleading due to the pervasive influence of state-driven values on their technology.
Chinese AI models are not expected to be entirely unbiased; indeed, every AI model possesses some biases, as Large Language Models are inherently not binary and encode values and morals. As a Harvard Law professor, Lawrence Lessig argues, code embodies social norms, regulations, and market forces that shape cyberspace.
China's approach to AI governance reflects lessons from the Cultural Revolution, Arab Spring, and Colour Revolutions, reinforcing its commitment to controlling cyberspace to prevent dissent and separatist movements, thereby influencing AI models.
China's AI models are no exception; they reflect state philosophy and communist ideology, necessitating vigilance among global users. China's approach to AI governance reflects lessons from the Cultural Revolution, Arab Spring, and Colour Revolutions, reinforcing its commitment to controlling cyberspace to prevent dissent and separatist movements, thereby influencing AI models.
DeepSeek has gained significant attention following President Donald Trump's characterisation of it as a "wake-up call," particularly affecting US companies like Nvidia in the stock market. President Donald Trump's warning about tariffs and trade war could further escalate tensions between the US and China across trade, technology, and governance domains. AI has emerged as a new battleground in US-China strategic competition, with DeepSeek being described as revolutionary and a "Sputnik Moment" that signals an intensifying AI race between the two powers.
China's positioning of AI as the new frontier of strategic competition is evident in its efforts to promote a Sino-centric digital order through technological influence, showcasing its civilisational advancement, a strategy similar to its approach with TikTok in "Project Texas." However, a contradiction arises between China's international rhetoric advocating "democratic AI" through platforms like TikTok and DeepSeek and its domestic imposition of strict digital controls.
This duality underscores Beijing's intent to gain international legitimacy while reshaping global internet norms to counter Western digital hegemony. Western concerns are further heightened by Chinese legal mandates, such as the Counter-Espionage Law 2014 and the 2017 National Intelligence Law, which compel private companies to cooperate with state intelligence agencies.
The growing demand for AI across infrastructural sectors has increased, posing significant threats to national security.
These regulations amplify fear over state-owned enterprises, particularly with tech companies and Chinese apps like TikTok, DeepSeek, and RedNotes (a browsing app). Meanwhile, China's advancements in EVs, drones, AI, and logistics are intensifying market competition, challenging US dominance, and aiming to isolate America. According to the AI Index Report 2024, China's AI innovations, patents, and startups, including those under Alibaba and DeepSeek, are reducing reliance on Silicon Valley. Additionally, DeepSeek's lower production costs are incentivising nations like India and Europe to accelerate their own AI investments, further fuelling this global technological race.
The growing demand for AI across infrastructural sectors has increased, posing significant threats to national security. This proliferation of AI and emerging technologies raises concerns that market competition and efficiency may overshadow democratic processes, prioritising utilitarian outcomes over genuine participatory democratic governance.
At the same time, the internet has fostered an age of superficial engagements, creating an illusion of transformation while substantive changes remain elusive. Zheng Yongnian, a prominent political scientist in contemporary China, explored these tensions in the book The Internet, State and Society in China (2007), highlighting China's preference for "Mr. Science" over "Mr. Democracy."
China's adherence to the principle of cyber sovereignty reinforces its digital authoritarianism, employing AI and critical emerging technologies for surveillance and social control and restricting external influence in the digital domain.
This raises a question: Is China truly interested in democratising the AI domain, or is its rapid AI development primarily driven by strategic competition with the US? As China transitions from a state of relative isolation to one of a global exhibition showcasing its civilisation prowess, it exacerbates challenges such as the digital divide, misinformation, and systemic bias. Meanwhile, security concerns suggest China's race to develop AI capabilities is not aimed at democratising the AI space but rather at strategic competition with the US.
Nistha Kumari Singh is a doctoral candidate and TMA Pai Fellow in the Department of Geopolitics and International Relations, Manipal Institute of Social Sciences, Humanities, and Arts, Manipal Academy of Higher Education (Institution of Eminence), Manipal, India.
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content -- blogs, longforms and interviews.