UN agency pushes global AI ethics norm that bans use of the technology for social scoring, mass surveillance purposes

Unesco’s AI guidelines oppose use of the technology for ‘invasive’ applications that ‘infringe on human rights and fundamental freedoms’
The UN agency’s recommendations come more than a month after China introduced its own set of ethical guidelines governing AI

Artificial intelligence

Coco Fengin Beijing

Published: 11:00pm, 25 Nov, 2021

Why you can trust SCMP

The world’s first international ethics guidelines for artificial intelligence, which ban the technology’s use for “social scoring or mass surveillance purposes”, have been adopted by a United Nations specialised agency that AI powerhouses the United States and Israel withdrew from in 2018.

The United Nations Educational, Scientific and Cultural Organisation (Unesco), which has 193 member states and is widely credited for protecting landmarks known as World Heritage Sites, said the guidelines serve as a global set of “recommendations” instead of a binding agreement, according to a statement on Thursday.

While major AI proponents, such as China, see the technology as a tool to help transform a broad swathe of industries, its various applications – from apps, social media and online retail to social scoring platforms and surveillance systems – have raised what Unesco describes as “fundamental ethical concerns” that may lead to “discrimination, inequality [and] digital divides”.

The 28-page Unesco document, officially known as the “Recommendation on the Ethics of Artificial Intelligence”, bans the use of AI for “social scoring or mass surveillance purposes” because “these types of technologies are invasive and they infringe on human rights and fundamental freedoms”, said Gabriela Ramos, the agency’s lead spokeswoman for the guidelines, without specifying any country or enterprise that promotes such activities.

Exterior view of the United Nations Educational, Scientific and Cultural Organisation headquarters in Paris, France. Photo: EPA-EFE

The guidelines call for “more transparency over the control of personal data” and “greater limits and awareness of the ability of AI to mimic human traits and behaviours”.

It wants to ensure that “the dominance of the English language in AI does not come at the expense of minority languages and broader cultural diversity and opinion”.

It also warns that “constant interaction with AI technology, including via social media algorithms, can negatively impact the mental health of both children and adults”, Unesco’s Ramos said.

Shenzhen is first Chinese city to draft regulations specifically for AI

Unesco’s recommendations come more than a month after China announced its own set of ethical guidelines governing AI, which focuses on protecting user rights and ways that align with the country’s goal to become a global AI leader by 2030.

Still, the UN agency’s AI guidelines encourage quite the opposite of what China has been doing with the technology. China currently has the world’s largest surveillance system, with 16 out of the top 20 most surveilled cities located in the country, according to a report published in May by British tech research firm Comparitech.

The recommendations from Unesco are not expected to “slow down China’s adoption of mass surveillance tools”, said Paul Bischoff, Comparitech’s editor and a regular commentator on cybersecurity and privacy. “China can always make a case, no matter how slight, for national security to support its surveillance systems.”

China this year already rolled out new data laws that would bolster protection of information in line with Unesco’s guidelines. These include the Data Security Law, which took effect in September, and the Personal Information Protection Law that was implemented earlier this month.

City governments across China have faced plenty of backlash against the increased use of AI-powered facial recognition systems. E-commerce hub Hangzhou, capital of eastern Zhejiang province, was the first city to pass a law that bans property management firms from forcing people to register their biometric identification, such as fingerprints, to enter a residential compound. This regulation will take effect in March next year.

While the US and close ally Israel have quit Unesco over certain policy differences, that has not stopped Washington from drawing up its agenda for international AI cooperation.

US and allies must set ‘democratic’ AI rules, Biden officials say

In July, senior White House officials appeared at a conference on emerging technologies to reinforce the message that Washington and its allies must close ranks to ensure that advances in artificial intelligence are developed in accordance with “democratic values”, and not left to China.

“We can’t let China write the rules around AI,” US Commerce Secretary Gina Raimondo told the Global Emerging Technology Summit in Washington.

Other international bodies have also been working on AI ethics. The Organisation for Economic Cooperation and Development, for example, published in 2019 the “Principles on Artificial Intelligence” that encourages “respect [for] human rights and democratic values” when using the technology.