In her new book, Empire of AI: Dreams and nightmares in Sam Altman’s OpenAI, Karen Hao writes about the Silicon Valley firms competing to be the first to reach artificial general intelligence and the damage they are inflicting on workers in poorer nations. 

The book comes at a time when OpenAI, Google, Meta and others are fast-tracking their AI projects amid rising competition from Chinese firms. Their models are built by a vast hidden workforce in nations such as Kenya, the Philippines and Venezuela that tag data, moderate content, and train and test AI models. 

These workers have complained of precarity and exploitation, including low wages, psychological trauma from moderating disturbing content, and being fired for unionizing. Tech companies are also accused of extracting the creative labor of artists whose copyrighted works are used to train large language models without their consent.

In her book, Hao — who was among the first journalists to gain access to OpenAI in 2019 — argues that this route to AI development is not the only way. As Pope Leo XIV warned this month of the challenges AI posed to human dignity, justice and labor, Hao spoke to Rest of World about the choices made by AI firms and alternative paths for AI development. 

The interview has been edited for length and clarity.

Data annotation firm Sama and its client Meta are facing a series of lawsuits from content moderators in Kenya and Ghana on exploitative working conditions. We’ve seen similar reports on OpenAI. What makes these workers vulnerable to the extractive practices you describe as “data colonialism”?

The countries where the digital workers live are still wrestling with the legacies of colonization. Many of them have underdeveloped political institutions and heavily volatile economies. When companies from the Global North enter these countries, they can easily find workers who are well-educated and have good internet connectivity—and are desperate for any job opportunity, and willing to work for any price. 

The governments of these countries are also in a tough position. Some don’t function well and are not acting in the best interest of their citizens. But even the ones that are trying to bring more economic opportunity have little power in stipulating the conditions under which foreign tech companies should operate. They’re willing to allow the companies in to secure the investment and be part of the global technology chain.

These labor practices happen so far away, and that’s by design. The companies choose places that are hidden from the consumers that they sell to.

Data annotation firms are often bound by legal contracts that limit what they can say, allowing Big Tech companies to distance themselves legally and ethically from their workers. What makes working conditions for these digital workers so challenging?

Content moderation for generative AI exposes workers to psychologically traumatic content. Regardless of the benefits they’re provided, whether it’s pay or mental health counseling, the work itself is extremely damaging to anyone who does it for long periods of time. It’s hard to imagine a version of this work that would not be harmful or debilitating.

AI companies have created this veneer of inevitability that, in order to have technological progress, workers will need to do this kind of labor. 

But that trade-off is false. Psychologically traumatizing work is not necessary for progress; it serves only a particular vision of AI development that uses colossal models trained on the whole internet, and that needs an army of workers to filter out its worst content.

Many research organizations outside Silicon Valley advance alternative notions of AI: well-scoped systems trained on well-curated datasets that do not encompass all of the content in the world. In that vision of AI, the need for psychologically compromising work could be significantly minimized or even eliminated.

Generative AI is replacing the clickwork that many migrants in countries like Venezuela relied on. What does that mean for these workers?    

The AI industry has had a long-standing ambition to replace clickworkers with AI models, but edge cases still require humans in the loop. Even so, the work remains precarious because the industry continually moves to different populations and geographies in search of cheaper, more compliant labor.

That’s why Venezuelan workers began falling off platforms [like Appen]. Improving economic conditions made the workers demand higher pay. 

Then, the industry shifted from image-based to language-based tasks, prompting companies to seek out English-speaking workers in Kenya, the Philippines, and India.

The AI industry shifts to different parts of the world whenever it encounters an obstacle. With so many countries undergoing economic or political crises, the labor pools for exploitation are abundant.

Artists, writers, and musicians in the U.S. have filed lawsuits against AI companies for training their models on copyrighted works. Writers and publishers in Singapore have rejected their government’s plan to train LLMs on their work, even as Big Tech firms lobby governments to allow use of copyrighted material under “fair use.” How does this impact artists whose work is used without consent or compensation?

I’ve spoken to artists who say that they do not have economic opportunities anymore. Concept artists, who generate visual ideas for film and advertising, have been especially hit hard. Karla Ortiz, one of the first concept artists who filed a lawsuit against generative AI companies, has described these jobs as solid middle-class jobs that you could build a life around. 

But now, concept artists are not getting jobs, or getting extremely underpaid jobs, because their work was taken by tech companies that created models to replace them. The fact that they’re rapidly losing economic opportunities shows definitively that AI-generated works are functioning as direct substitutes of their original work.

What are some examples of alternative visions for AI that can produce ethically sourced systems—for those who create the data and the workers who train the models?

We can develop AI systems with data that are ethically sourced, labeled, and curated. 

A Maori media organization in New Zealand developed a speech recognition tool for the Maori language with data given by their community members, with consent. They also developed a license requiring any organization seeking to access their data to sign a legally binding agreement that governs how they will use, protect, and secure the data. This ensures the data is not fed into applications that might ultimately harm the Maori community.

In generative AI, the BigScience initiative was an open-source effort involving hundreds of scientists around the world to develop a large language model using archival data from governments and public institutions and donated data. There are a lot of these kinds of models out there. But tech companies have chosen not to adopt any of them. And they have been so effective in completely hijacking the public’s imagination for what AI should look like and how it should be developed.

As Jessy Wu prepared to launch her Beijing-based artificial intelligence startup overseas last year, she grappled with a question familiar to many Chinese tech founders: Should her company hide its Chinese identity?  

For years, as governments in the U.S., Europe, and elsewhere restricted Chinese apps and technologies over national security concerns, many companies practiced what came to be known as China shedding — downplaying and obscuring their origin in overseas markets to win over investors and customers. But the global success of Chinese social media apps and the launch of a successful AI model by startup DeepSeek earlier this year have convinced some founders like Wu that they no longer need to do so. 

“DeepSeek has shown us that as long as our product is competitive, we should not worry too much about being a Chinese company,” Wu told Rest of World. “The popularity that TikTok and RedNote have been receiving in the U.S. shows us that users are actually welcoming towards good Chinese products,” Wu said, adding that her film-generation platform has steadily grown its user base across the U.S. and East Asia while being openly Chinese.

Wu’s confidence signals a shift among Chinese startups. Rather than avoiding calling attention to their Chinese roots, they are promoting the competitive advantages of being Chinese. DeepSeek’s rise has attracted foreign investors searching for the next Chinese tech breakout, and Chinese tech companies are launching English-language product videos and even speaking with foreign media — something that executives had shied away from.

Startup founders who spoke with Rest of World said they are leaning into the strengths of their Chinese roots — such as engineering talent and cheaper production costs — while hiring foreign staff and complying with local regulations to avoid the sort of scrutiny faced by larger firms, such as TikTok, drone company DJI, and top electric vehicle maker, BYDiBYDBYD Auto is a Chinese carmaker that became the world’s leading EV manufacturer in 2023, competing with Tesla for market share and global attention.READ MORE.

Jia Zijian, who launched his language-learning app TalkMe on global app stores last year, said DeepSeek’s success is proof of Chinese startups’ competitiveness.

“DeepSeek gives Chinese tech startups an image boost. … I believe foreign companies can hardly compete with the kind of innovation we create with such cost efficiency,” Jia told Rest of World.  

Although his 10-person team is based in Beijing, Jia registered his company in Singapore last year, like many Chinese companies have done to avoid close scrutiny, a move that has come to be known as Singapore-washing. His goal was to better attract foreign capital, he said.

“Where we place our company tells people where we are focused,” he said. Conversations with investors led him to believe that “there is no need to cover up, but also no point branding as Chinese,” he said, adding that his app was built entirely on foreign AI models and that all transactions were made through foreign banks.

Jia’s strategy is a common one. Some Chinese apps that launched in the past three years, such as short-form video app ReelShort and AI video generator HeyGen, have moved their operations and staff to California even as they have been open about their Chinese origin.

This approach is a mix of pride and pragmatism, said Ivy Yang, a China tech analyst and founder of New York–based consulting firm Wavelet Strategy, which provides communications and reputation management services for Chinese companies in the U.S.

“I wouldn’t go as far as to say that Chinese AI companies will suddenly make their Chineseness a selling point, but [it was] definitely a sigh of relief [after DeepSeek’s launch] that they will be less likely penalized for it,” Yang told Rest of World. 

Chinese companies have long hired Westerners to project an international image — sometimes as executives, other times as paid actors for events — a practice so widespread it is known as the rent-a-foreigner industry. That strategy has since evolved. 

Even a year or two ago, conversations with Chinese clients always began with “How do we eliminate the Chinese aspect of our image?” Chris Pereira, Singapore-based chief executive of iMpact, a communications firm advising Chinese companies expanding abroad, told Rest of World. 

Now, Pereira said, more clients are asking how to tailor their products and services for overseas markets. “It is not to hide their identity, it is to localize their operations,” he said. 

“DeepSeek and [other] successful Chinese companies overseas right now are giving the people coming in the wave behind them more confidence to just be themselves — it’s OK to be Chinese overseas.”

For startups seeking foreign investment, registering overseas and setting up separate business operations outside of mainland China is a natural choice, Bob Chen, an economist at China-based investment firm FG Venture, told Rest of World. 

“Whose money a startup is trying to get would determine how exposed they want to be about their Chinese affiliation,” Chen said. 

Chinese startup Butterfly Effect, creator of general-purpose AI agent Manus, is reportedly considering relocating its headquarters to Singapore and has already registered a corporate entity there, as it raised $75 million from U.S. venture capital firm Benchmark. 

Chinese tech companies remain wary of political tensions and regulatory scrutiny, particularly in the U.S. TikTok faces a potential ban in June unless its Chinese parent company ByteDance sells it, although U.S. President Donald Trump has said he would extend the deadline. The short-video app, registered in Singapore, has faced concerns regarding data security and privacy as well as Chinese government influence. TikTok, which has a Singaporean CEO, has denied these allegations. 

The U.S. government is also assessing whether Chinese drone-maker DJI should face restrictions on national security grounds. BYD faces similar concerns in some countries.

More Chinese tech companies are placing greater emphasis on compliance with local regulations on data security and technology transfer to avoid scrutiny, according to a study by consulting firm Deloitte. 

As Chinese companies expand abroad, they will focus more on creating distinctive global brands, Pereira said. Companies going overseas are choosing Western names, he said, citing Temu, the international sister brand of popular shopping site Pinduoduo in China, as an example. 

“It’s just like Coca-Cola has a Chinese name, Kekoukele. … if we look at it from that perspective, it’s exactly what the American brands do globally too,” he said. “It is not to hide their identity, it is to localize their operations.”