China’s technology group Tencent is making a push on the international stage for its vision of ethics in artificial intelligence in a break with the company’s traditionally reclusive style.
In a rare interview, China’s most valuable publicly listed tech company said it favoured an approach to ethics that combined not only socially beneficial uses of AI for medical purposes or agriculture but also ensured a “social contract” between companies and users to govern the use of personal data.
“As our company gets larger in significance in our position in society, we should begin to question what’s the soul of the company,” Seng Yee Lau, Tencent’s senior vice-president and chairman of global branding and group marketing, told the Financial Times.
Mr Lau said Tencent summed up its vision on ethics with the slogan “tech for good”, a notion rooted in Chinese philosophies and the idea of good as coming from within the individual, rather than being enforced by governments.
Tencent’s push comes at the same time western tech companies such as Google have tried to set up AI ethics boards. However, some employees have criticised them as lacking teeth, and while tech giants may invest in security to keep user data safe, many have been forced by their governments to hand over data — with Tencent being no exception.
“Tencent created its AI for Social Good platform in January 2018, before Google launched a programme with the same name,” said Danit Gal of the Cyber Civilization Research Center at Keio University in Tokyo. “They’re really not that far behind and have probably done more on this topic than other companies.”
Tencent, China’s most valuable publicly listed tech company, runs WeChat, the messaging and payment platform that forms the bedrock of China’s internet.
That also means Tencent takes the brunt of the Chinese government’s demands — implicit and explicit — to censor internet content in what advocacy group Freedom House ranks as the world’s worst country for internet freedoms.
The WeChat app automatically censors keywords deemed politically sensitive, while in-house censors also delete posts and accounts.
Mr Lau praised the EU’s General Data Protection Regulation, saying it “was a huge inspiration” for Tencent. “Billions of users have entrusted us with their personal sensitive information, this is the reason we must uphold our integrity above the requirements of the law,” he said.
“As a Chinese company starting from a social network business, we actually care more about user privacy,” Mr Lau added, describing how the company has its own team of hackers dedicated to finding vulnerabilities in Tencent’s software. Tencent’s team, he said, has helped Tesla test its systems.
But such tests cannot defend against government requests for user data. Beijing’s crackdown on internet discourse has grown in recent years. In one example, a man was sentenced to two years in jail for mocking Xi Jinping in a private WeChat message.
Mr Lau said that he hoped to work with an international alliance of academics, companies, governments and multinational organisations, such as the UN, in forming ethical approaches to new technologies.
“China is actually seen as a role model,” he said, mentioning his meetings with the president of France and the prime ministers of Japan and Malaysia. “You realise the reason they were willing to come talk to you is because there was something they’d like to learn from you.”
One key difference between Tencent and western companies’ approaches to AI ethics, Ms Gal added, was that Tencent’s approach “involves a variety of stakeholders like public interest groups, universities, even Buddhist monks”.
Mr Lau said he wanted to move beyond the term “corporate social responsibility”, saying that “it assumes that I have to please the board, donate some money. That’s not right — if you are really in charge of social co-operative responsibility, then you have to study the UN 2030 agenda for sustainable development,” he said.
Speaking at an event in Dubai on Wednesday, Mr Lau focused on the UN’s 17 sustainable development goals, such as ending extreme poverty. In talking about AI ethics, Chinese companies have so far focused more on concrete achievements — such as facial-recognition programs that can help track missing children — rather than general principles.
As examples of ethical achievements in AI, Mr Lau spoke of Tencent’s medical imaging procedures, which he said assist with the diagnosis of glandular cancer and colorectal polyps with 97 per cent accuracy.
Tencent has also invested in Phytech, a farming company that uses data to optimise crop cultivation, Mr Lau said.