Since its roll-out in January
2025, Chinese generative Artificial Intelligence (AI), DeepSeek R1, has
challenged existing assumptions about AI development, cost efficiency, and
global competition, with some even going so far as to claim that it has levelled
the playing field of AI development. As the Chinese Communist Party (CCP)
promotes DeepSeek abroad, making cutting-edge AI technology accessible to the
Global South, this comes with hidden risks and human-rights-related
consequences.
Previously, experts had believed that cutting-edge AI innovation would take place in the wealthiest nations, furthered by powerful corporations due to the related immense hardware costs. Even Sam Altman, OpenAI’s founder, stated during a visit to India back in 2023 that training foundational AI systems on a USD 10 million budget would be futile: “You can try, but it would be hopeless to compete with us.”
Yet, Altman’s comments have been upended by DeepSeek’s breakthrough, and he had to admit that the Chinese AI “is an impressive model, particularly around what they're able to deliver for the price.”
The made-in-China AI has sparked interest across developing countries and paved the way for China to position itself as a key tech partner for the Global South, pushing the narrative of bridging the AI gap for less developed regions.
In addition, the strategic use of AI for foreign relations, through initiatives like the Digital Silk Road and the proposal of establishing the Global AI Cooperation Organization, could significantly reshape regional power structures in China’s favour.
DeepSeek R1’s open-source nature means developers can download, modify, and fine-tune the model to suit local conditions. This could be adapted in regional languages such as Swahili, Hausa, or Amharic, or to the unique economic landscapes.
The Indonesian government, for example, is planning to develop its own local model of DeepSeek for use in the government and private sectors.
These ties not only enable technological growth but also foster geopolitical alignment.
Some analysts believe that a deepening dependence on Chinese tech, standards, and data flows in Africa and Asia shifts the power dynamic away from the US and Europe. It could also influence the balance in multilateral forums and international bodies, such as the United Nations.
Shaping global narratives
However, DeepSeek, whether used abroad or domestically, aligns its chatbot outputs with the CCP's official viewpoints. While several news outlets have published reports about the censoring or altering of responses on politically sensitive topics, the real danger lies not with the simple withholding or removal of information.
According to Alex Colville, a researcher at the China Media Project (CMP), the shaping of narratives through generative AI is a sophisticated manipulation of public opinion. Below is an example given in the CMP’s report:
“When asking questions about natural disasters in China, for example, the model treats Chinese government data and sources as infallible, and portrays the leadership’s response as being effective, transparent and humane. Meanwhile, dissenting voices are either minimized, omitted entirely, or explained away as “biased” or lacking understanding.”
In other words, DeepSeek’s large language model (LLM) could become a conduit for Beijing’s political narratives and shape narratives on sensitive issues like human rights, Taiwan, economic development, and its role in global trade.
Worse still, the Chinese political narratives can be exported to other countries through the adaptation of the LLM. CMP's researcher warned:
But if even cutting-edge tech companies in developed nations, for all their resources and funds, are struggling to train propaganda out of DeepSeek, what hope do start-ups in the Global South have?
DeepSeek has raised concerns in countries like Australia, South Korea, and Taiwan, which have banned the use of DeepSeek on government devices, citing security risks and a lack of transparency as to how DeepSeek processes users’ personal information.
Beyond DeepSeek, China’s AI-driven propaganda
Several reports showed that China has utilized DeepSeek and other AI technologies to create and disseminate propaganda, including deepfakes and manipulated content, to influence public opinion and sow discord in other countries.
Data security experts, Brett Goldstein and Brett Benson of Vanderbilt University, pointed to a mainland Chinese AI-powered company, Golaxy, which, in addition to monitoring public opinion, also uses AI to collect massive amounts of data from Chinese and Western social media platforms and create “humanlike bot networks” and “psychological profiling to target individuals.”
While the company denied the above claims, the researchers warned in an opinion piece on the New York Times:
A.I.-driven propaganda is no longer a hypothetical future threat. It is operational, sophisticated and already reshaping how public opinion can be manipulated on a large scale.
The leaked documents show that the company had operations in Hong Kong in 2020 as the city passed its national security law by “identifying thousands of participants and thought leaders from 180,000 Hong Kong Twitter accounts” and then “went after what it perceived as lies and misconceptions” via its bot networks. A similar operation was detected during Taiwan’s 2024 general elections.
The researcher duo explained how the AI-powered propaganda had manipulated public opinion en masse on social media:
By extracting user data and studying broader patterns, A.I. can build synthetic messaging designed to appeal to a wide spectrum of the public. It can adapt to a user’s tone, values, habits and interests, according to the documents. Then it can mimic real users by liking posts, leaving comments and pushing targeted content.
A 2024 report, “Same Targets, New Playbooks,” released by Microsoft Threat Intelligence, also affirmed how China had powered up its online influence campaigns in the Asia-Pacific region with AI-generated content. The report said:
The influence actors behind these campaigns have shown a willingness to both amplify AI-generated media that benefits their strategic narratives, as well as create their own video, memes, and audio content. Such tactics have been used in campaigns stoking divisions within the United States and exacerbating rifts in the Asia-Pacific region — including Taiwan, Japan, and South Korea.
Export of surveillance technology
Alongside LLMs, China has also promoted the export of other AI technology, such as facial recognition. Some have raised concerns over the impact of the surveillance technology. Law professor, Anu Bradford, for example, expressed her concern in her study of China’s digital authoritarianism that through exportation of digital technology and its infrastructure, the Chinese government will promote its’ regulatory standards and surveillance practices across the world:
Many receiving countries have welcomed Chinese technologies and accompanying regulatory standards as a path toward digital sovereignty and development. For authoritarian governments, an additional motivation has been to gain access to surveillance technologies that they eagerly use toward illiberal ends.
For developing countries, China’s cheap surveillance technology could address many problems, such as cracking down on crime, but more research is needed to see its impact on democratic institutions, civil liberties, and global power dynamics, as public affairs expert Sheena Chestnut Greitens said.
Beijing is positioning itself as a key tech partner for the Global South by offering low-cost, adaptable systems such as DeepSeek R1. Through such tactics, the authoritarian state is building not only digital capacity but also potentially political alignment by positioning itself as the leader of global AI governance.
At present, the US is racing with China in AI governance, mainly to protect its global economic interests. Meanwhile, civil society's calls to develop a human rights framework in global AI governance have remained sidelined. Without adequate intervention, we could see a worldwide normalization of AI technology for surveillance and propaganda purposes.
Previously, experts had believed that cutting-edge AI innovation would take place in the wealthiest nations, furthered by powerful corporations due to the related immense hardware costs. Even Sam Altman, OpenAI’s founder, stated during a visit to India back in 2023 that training foundational AI systems on a USD 10 million budget would be futile: “You can try, but it would be hopeless to compete with us.”
Yet, Altman’s comments have been upended by DeepSeek’s breakthrough, and he had to admit that the Chinese AI “is an impressive model, particularly around what they're able to deliver for the price.”
The made-in-China AI has sparked interest across developing countries and paved the way for China to position itself as a key tech partner for the Global South, pushing the narrative of bridging the AI gap for less developed regions.
In addition, the strategic use of AI for foreign relations, through initiatives like the Digital Silk Road and the proposal of establishing the Global AI Cooperation Organization, could significantly reshape regional power structures in China’s favour.
DeepSeek R1’s open-source nature means developers can download, modify, and fine-tune the model to suit local conditions. This could be adapted in regional languages such as Swahili, Hausa, or Amharic, or to the unique economic landscapes.
The Indonesian government, for example, is planning to develop its own local model of DeepSeek for use in the government and private sectors.
These ties not only enable technological growth but also foster geopolitical alignment.
Some analysts believe that a deepening dependence on Chinese tech, standards, and data flows in Africa and Asia shifts the power dynamic away from the US and Europe. It could also influence the balance in multilateral forums and international bodies, such as the United Nations.
Shaping global narratives
However, DeepSeek, whether used abroad or domestically, aligns its chatbot outputs with the CCP's official viewpoints. While several news outlets have published reports about the censoring or altering of responses on politically sensitive topics, the real danger lies not with the simple withholding or removal of information.
According to Alex Colville, a researcher at the China Media Project (CMP), the shaping of narratives through generative AI is a sophisticated manipulation of public opinion. Below is an example given in the CMP’s report:
“When asking questions about natural disasters in China, for example, the model treats Chinese government data and sources as infallible, and portrays the leadership’s response as being effective, transparent and humane. Meanwhile, dissenting voices are either minimized, omitted entirely, or explained away as “biased” or lacking understanding.”
In other words, DeepSeek’s large language model (LLM) could become a conduit for Beijing’s political narratives and shape narratives on sensitive issues like human rights, Taiwan, economic development, and its role in global trade.
Worse still, the Chinese political narratives can be exported to other countries through the adaptation of the LLM. CMP's researcher warned:
But if even cutting-edge tech companies in developed nations, for all their resources and funds, are struggling to train propaganda out of DeepSeek, what hope do start-ups in the Global South have?
DeepSeek has raised concerns in countries like Australia, South Korea, and Taiwan, which have banned the use of DeepSeek on government devices, citing security risks and a lack of transparency as to how DeepSeek processes users’ personal information.
Beyond DeepSeek, China’s AI-driven propaganda
Several reports showed that China has utilized DeepSeek and other AI technologies to create and disseminate propaganda, including deepfakes and manipulated content, to influence public opinion and sow discord in other countries.
Data security experts, Brett Goldstein and Brett Benson of Vanderbilt University, pointed to a mainland Chinese AI-powered company, Golaxy, which, in addition to monitoring public opinion, also uses AI to collect massive amounts of data from Chinese and Western social media platforms and create “humanlike bot networks” and “psychological profiling to target individuals.”
While the company denied the above claims, the researchers warned in an opinion piece on the New York Times:
A.I.-driven propaganda is no longer a hypothetical future threat. It is operational, sophisticated and already reshaping how public opinion can be manipulated on a large scale.
The leaked documents show that the company had operations in Hong Kong in 2020 as the city passed its national security law by “identifying thousands of participants and thought leaders from 180,000 Hong Kong Twitter accounts” and then “went after what it perceived as lies and misconceptions” via its bot networks. A similar operation was detected during Taiwan’s 2024 general elections.
The researcher duo explained how the AI-powered propaganda had manipulated public opinion en masse on social media:
By extracting user data and studying broader patterns, A.I. can build synthetic messaging designed to appeal to a wide spectrum of the public. It can adapt to a user’s tone, values, habits and interests, according to the documents. Then it can mimic real users by liking posts, leaving comments and pushing targeted content.
A 2024 report, “Same Targets, New Playbooks,” released by Microsoft Threat Intelligence, also affirmed how China had powered up its online influence campaigns in the Asia-Pacific region with AI-generated content. The report said:
The influence actors behind these campaigns have shown a willingness to both amplify AI-generated media that benefits their strategic narratives, as well as create their own video, memes, and audio content. Such tactics have been used in campaigns stoking divisions within the United States and exacerbating rifts in the Asia-Pacific region — including Taiwan, Japan, and South Korea.
Export of surveillance technology
Alongside LLMs, China has also promoted the export of other AI technology, such as facial recognition. Some have raised concerns over the impact of the surveillance technology. Law professor, Anu Bradford, for example, expressed her concern in her study of China’s digital authoritarianism that through exportation of digital technology and its infrastructure, the Chinese government will promote its’ regulatory standards and surveillance practices across the world:
Many receiving countries have welcomed Chinese technologies and accompanying regulatory standards as a path toward digital sovereignty and development. For authoritarian governments, an additional motivation has been to gain access to surveillance technologies that they eagerly use toward illiberal ends.
For developing countries, China’s cheap surveillance technology could address many problems, such as cracking down on crime, but more research is needed to see its impact on democratic institutions, civil liberties, and global power dynamics, as public affairs expert Sheena Chestnut Greitens said.
Beijing is positioning itself as a key tech partner for the Global South by offering low-cost, adaptable systems such as DeepSeek R1. Through such tactics, the authoritarian state is building not only digital capacity but also potentially political alignment by positioning itself as the leader of global AI governance.
At present, the US is racing with China in AI governance, mainly to protect its global economic interests. Meanwhile, civil society's calls to develop a human rights framework in global AI governance have remained sidelined. Without adequate intervention, we could see a worldwide normalization of AI technology for surveillance and propaganda purposes.
No comments:
Post a Comment