June 13, 2023
Geoffrey Cain, Senior Fellow, Foundation for American Innovation
“America, the Vanguard of Democracy, Must Stand Up to China’s AI Totalitarianism”
Chairman Ossoff, Ranking Member Blackburn, and members of the Subcommittee:
Thank you for the opportunity to testify today. My testimony has two purposes:
First, to outline how China has created the world’s most sophisticated and terrifying
surveillance state using novel artificial intelligence (AI) technologies, and how American
business elites helped make this happen.
Second, to suggest ways that the US can defend the use of AI with respect to democracy
and human rights, to ensure that the Chinese Communist Party (CCP) cannot advance its
malign global agenda with AI tools.
With AI, America’s elites have learned little about the perils of engaging with China’s
one-party authoritarian state.
On Friday, OpenAI CEO Sam Altman dialed into the annual conference at the Beijing
Academy of Artificial Intelligence, three weeks after he testified before another
subcommittee here at the Senate Judiciary Committee. He called on the People’s Republic
of China—a one-party authoritarian state that has used AI to carry out genocide against an
ethnic minority—to help shape global AI safety guardrails. “With the emergence of
increasingly powerful AI systems,” he said, “the stakes for global cooperation have never
been higher.”
To anyone who’s lived in China, this was a curious and mind-boggling call to action. The
Chinese Communist Party (CCP) has engineered a vast AI-powered surveillance system
literally called “Sky Net.” It runs AI-powered “alarms” that notify the police and
intelligence services when someone unfurls a banner, when a foreign journalist is
traveling to certain parts of the country, and when someone from an ethnic minority is present.
The government accuses entire groups, such as Muslim Uyghurs, of posing a
terrorist threat, and relentlessly persecutes them with the use of AI tools.
It sounds like a dystopian science fiction story—think 1984 or Minority Report—but the
CCP’s AI totalitarianism has become a fact of daily life for the more than 1.4 billion people
in China. In fact, the Chinese technologists who spoke at the same conference as Mr.
Altman were some of the very people who built this monstrosity. They were executives at
iFlyTek and Huawei, two AI giants and are heavily sanctioned by the US government for
their involvement in human rights abuses.
If Mr. Altman plans on cooperating with China’s AI developers, he better figure out who he’s working with.
I’ve witnessed the results of their work firsthand. As an investigative journalist formerly in
China, I was among the first people to document and expose the horrific surveillance state
that oppressed the Uyghur population in the far western region of Xinjiang. Since 2017,
the atrocity has morphed into the largest internment of ethnic minorities since the
Holocaust, which the US State Department calls a genocide.
Chinese authorities have hauled away 1.8 million people to concentration camps—about
one-tenth of the ethnic minority population in Xinjiang—and have forced many of them
into slave labor.
Because they have read too many books or have been caught praying,
they have been declared enemies of the state, despite not being formally charged with any
crime. This was all with the help of the AI surveillance system that scooped up data from
facial recognition, voice recognition, and a network of police cameras covering every
possible square inch of the region. Party authorities told Uyghurs they wanted to “cleanse”
their minds of what they called “ideological viruses.”
In December 2017, I was kicked out of China while researching my book, The Perfect Police
State: An Undercover Odyssey into China’s Terrifying Surveillance Dystopia of the Future. Ever
since then, the AI-fueled police state has expanded to alarming levels. In 2018, I moved to
Turkey and, for three years, tracked down former intelligence officers from China’s
Ministry of State Security, the powerful and secretive intelligence body. They had helped
set up the AI surveillance systems in Xinjiang, were targeted by those same systems
because they were Uyghurs, and then defected to safety.
These intelligence officers drew detailed diagrams in my possession that showed the
workings of these surveillance systems and how facial recognition and voice recognition
technologies helped fuel them. What they revealed was alarming, but not surprising. The
highest echelons of CCP leadership held centralized control over many AI surveillance
systems, as well as direct lines of influence over Chinese mega-companies such as Huawei
and ByteDance. With the help of these companies, China’s government had been making a
concerted, malicious effort to expand these surveillance capabilities all over the world.
The development of AI is at the heart of China’s global ambitions
The surveillance state that began in Xinjiang was a taste of the horrific power of AI when
placed in the wrong hands. “Advanced technology is the sharp weapon of the modern
state,” China’s President Xi Jinping said in a 2013 speech. In July 2017, China unveiled its National AI Development Plan, calling AI a “historic opportunity” and pledging to align
developments in AI with the government’s authoritarian values. China has declared its
goal as becoming the world leader in AI by 2030.
The goal reflects the totalitarian ambitions of President Xi, who has led the efforts to clamp down Uyghurs, Tibetans, Mongolians, and religious and political dissidents of all stripes.
Since then, we’ve seen the expansion of China’s technology companies, using AI and other
novel developments, all over the world. Huawei, the heavily sanctioned telecommunications firm, has led efforts to establish global surveillance systems, usually under the guise of AI-powered “smart cities” designed to fight crime and regulate traffic, but that in reality have been used to equip governments with the tools to spy on political dissidents. In October 2022, the FBI arrested two Chinese nationals who stood accused of bribing an undercover FBI officer to obtain inside intelligence about an investigation into Huawei.
Meanwhile, ByteDance, the $220 billion mega-firm that owns TikTok, stands accused by a
whistleblower of running an in-house CCP Committee that had access to all the app’s data,
including data stored in the US, according to a court filing. Other sanctioned, lesser known firms, such as AI facial and voice recognition companies iFlyTek, SenseTime, and Megvii, have emerged as global billion-dollar unicorns with the backing of the Chinese state and the involvement of US venture capital funds.
This situation is proving hard to continue in the age of technological decoupling. This
month, Sequoia Capital, the preeminent venture capital firm that originally invested in
Apple and Facebook, announced that it was splitting off its Chinese arm into a separate
company. Sequoia’s China business was core to helping build China’s AI industry, with a
reported $22 billion stake in ByteDance, to name one of many examples. Sequoia’s spinoff suggests that American business executives are waking up to the unavoidable risks of doing business in China—of inadvertently helping build China’s AI systems that damage human rights and the public good.
Generative AI is a threat to CCP censorship
In April 2023, the Cyberspace Administration of China announced draft regulations for
generative AI, setting down potential rules that chatbot-produced content follow “socialist
core values” and avoid information that undermines “state unity.” The CCP’s goal is a
continuation of its past strategy to align new technologies and censor information in line
with its political values. ChatGPT has not made its service available in China, but there is
already significant demand. The black market is already flourishing with offerings of
overseas ChatGPT access to people in China, but these days could be numbered.
Generative AI, however, is a departure from the surveillance technologies that have
defined the evolution of China’s political censorship. Generative AI services have the
potential to empower regular people who want to produce large amounts of content that
challenge government propaganda and narratives. The question is whether China’s “Great
Firewall”—the harsh internet censorship system—can stand up to the potential of
generative AI. Will China one day see an information renaissance, with stories of the
Tiananmen Square massacre and Hong Kong protestors spread across the internet
through uncontrollable chatbots?
Given the CCP’s enormous success at censorship so far, I believe that it will once again
succeed in coercing and coopting Chinese technology firms and transforming generative
AI into a tool of state oppression. American technologists will unwittingly assist CCP goals
if they cooperate too eagerly with state-connected Chinese companies, institutes, and
people. As we have learned over the last decade, this is the sad truth of being a
technologist in China.
The US must use its global technological leadership to protect democracy and human
rights from China’s AI threats
The CCP is the greatest threat to human rights and democracy around the world. Although
China is quickly catching up to US innovation, the US remains the leader in AI
development. We must abandon the misguided idealism of working with Chinese
companies and government bodies with the hope that AI will change the political system,
allow for the opening of democratic discourse, and create safer global AI regulations.
Rather than helping advance innovation, we will be doing the world a disservice by
handing the keys to the CCP. Under Chinese law, these advanced AI applications will
inevitably be used to oppress human rights and expand China’s authoritarian footprint.
Rather, we should use our position of strength and our democratic values to carry out a
two-fold strategy. First, AI talent and innovation must flow towards the direction of
America and its allies. We must influence global AI standards, attract global AI talent away
from China, and secure our software and hardware ecosystems from China’s malign
influences. Second, the most advanced American technologies and investments must not
be allowed to flow in the direction of China. We must work against China’s ambitions to
develop advanced AI systems, influence global standards, and oppress dissidents around
the world. The specific policy steps are as follows:
The US must take the lead in developing global AI standards that uphold human
rights and democratic values.
The CCP has loudly used multilateral membership bodies—the United Nations, the
World Health Organization, and so forth—to shape global technology and science
standards in its interests and to make countries all over the world dependent on
Chinese technological innovation. The US must not shirk its global leadership,
which would mean ceding ground to China and abandoning our allies in a moment
of global struggle.
In November 2021, 193 countries adopted the first-ever global agreement of AI
ethics under the United Nations Educational, Scientific, and Cultural Organization
(UNESCO), calling for a “do no harm” principle, personal data protection, and
measures to prevent fairness and non-discrimination.16 The US should leverage
other United Nations bodies and the International Organization for Standardization
(ISO) to build democratic AI principles and ensure that China’s authoritarian goals
do not crush the principles of human rights.
American companies that help build China’s oppressive AI ecosystem must be
held accountable.
China built its AI surveillance apparatus with the connivance and complacency of
major American technology firms. The science corporation Thermo Fisher, for
example, was caught selling DNA collection equipment directly to Xinjiang police
authorities who used them for mass gathering of genetic data on the minority
Uyghur population. Since the late 1990s, Microsoft has established itself as the
training ground for China’s AI elites through its Beijing-based laboratory, Microsoft
Research Asia. The laboratory has trained many of the AI leaders and developers
who went on to found or join the executive leadership of rights-abusing firms such
as Sensetime, Megvii, and iFlyTek. Beginning in 2019, the US government has
sanctioned these individuals and their companies.
So far, American technology giants have faced no punishment for their
involvement in China’s surveillance state. This subcommittee may consider
drafting a bill that requires public corporations to publish their due diligence
reports on their activities in China and the risks they have encountered with
regards to human rights there. The subcommittee may also consider drafting a bill
that criminalizes specific American business activities in China that are likely to
support, directly or indirectly, human rights abuses by the CCP. This would include
prison time for American business executives involved helping develop any form of
AI in partnership with a Chinese entity, if the CCP will likely use that technology for
the oppression of human rights and democratic values.
Because Chinese software companies are required to partake in Chinese state
intelligence operations, they should be compelled to separate their American
businesses.
Over the past decade, China has enacted a raft of draconian laws, such as the
National Security Law and the National Intelligence Law, that require people in
China to assist the government in intelligence-gathering when called upon, among
other requirements. While we in America have a system of due process and
checks and balances that can guard against data overreach, in China no such rights
exist. The private and personal data of Americans is not safe in the hands of
Chinese-owned apps such as TikTok and Temu, whose owners and employees in
China are required to hand over data to the state if it’s requested.
Apps like TikTok are beginning to form the core of the US information
environment, with sophisticated algorithms that recommend highly addictive content, while being used to spy on US citizens. This is a gaping breach of our ability to protect democratic values and human rights here in the US. In the event of conflict with China—an increasing likelihood with China’s aggressive military posture—these apps have the potential to become misinformation machines
designed to manipulate Americans with sophisticated and algorithmic propaganda.
The solution is to force these firms to spin off their American operations into
separate companies, ensuring their safety from CPP meddling.
America and its allies must secure and coordinate global supply chains for
advanced AI logic chips.
The US has made remarkable progress in legislating and implementing export
controls that prevent American firms from selling advanced chips and their
components to China. In October 2022, the Biden administration implemented the
most recent round of sanctions, restricting the export of certain services and
equipment to China, effectively placing China generations behind American chip
technologies for the latest AI applications. Four months later, in February 2023,
the Department of Commerce opened the first round of company grants under the
CHIPS and Science Act, hoping to reshore semiconductor manufacturing
capabilities and make the US more self-sufficient.
The CHIPS and Science Act, however, is the starting point and not the last step.
Advanced semiconductors are the most complex devices that humankind has ever
made—and they cannot simply be manufactured end-to-end in the US. Chip supply
chains depend on thousands of suppliers all over the world. The US needs to better
coordinate with its key chip-producing and component-producing partners—South
Korea, Taiwan, Japan, and the Netherlands—by upgrading the “Chip 4” talks into a
formal consortium for coordinating R&D innovations.
The upgrade will enhance the implementation of the CHIPS and Science Act and
the future of AI technologies by adding an element of multilateralism. Our
technological partners will have better reason to believe their contributions to the
US manufacturing ecosystem are profitable and worthwhile, a hedge against CCP
aggression. If we can form a true semiconductor alliance, China will be unable to
bully individual countries into supplying critical chip technologies for its AI
systems.
As we enter the unprecedented age of generative AI, we must not allow China, a one-party
authoritarian state, to infect the global AI ecosystem where it will oppress human dignity,
civil liberties, and rule of law. We have seen the CCP’s willingness to carry out genocide
against its people with the help of AI surveillance systems. Now we must find ways to
ensure that the words “never again” hold true. Thank you, Senators, for having me here
today. I look forward to answering your questions.
Read the original publication here.