The most AI-eager country in Southeast Asia is burning out because of it

Vietnam uses AI more than any of its Southeast Asian neighbors, and a growing number of its office workers are now discovering the cost: a form of mental exhaustion researchers call "AI brain fry."

Minh Thu, who works on a content team at a large corporation in Hanoi, has been paying with her sleep and her health. A year ago her team had six people; today it has three.

Her manager set the new terms bluntly: AI now handles 70% of the basic work, so the people who remain must deliver the other 30% at an exceptional level "or leave," she recalled. To prove she was "still worth more than an algorithm," she stretched her days from eight hours to 14.

"Every time my boss praises a new AI tool in the group chat, I push myself to work more and faster," she said. The fear of error is constant. "If I make a mistake, the company docks it straight from my salary and my performance review, so I can't eat or sleep properly."

Hoang Long, a 30-year-old communications worker at another firm, hit the wall harder. When his company brought in AI to generate content, management cut project deadlines from two weeks to five days on the logic that "AI is helping now."

Running three or four tools at once to write copy, design graphics and check data, he collapsed at the office in mid-May from low blood pressure and exhaustion, and a doctor ordered him to take leave.

Several generative AI apps on a smartphone: Copilot, DeepSeek, Gemini, AI Hay, ChatGPT and Grok. Photo by VnExpress/Luu Quy

Thu and Long are not isolated cases. In its e-Conomy SEA 2025 report, released in November 2025, Google, Temasek and Bain & Company found that 81% of Vietnamese users interact with AI tools daily, the highest rate in Southeast Asia, narrowly ahead of Indonesia's 80% and well above Thailand's 74%.

The same report put Vietnam first in the region on two further measures of enthusiasm: 83% of users are actively learning or upskilling in AI, and 96% say they are willing to share data with AI agents.

The intensity shows up in app data too. Vietnamese users averaged 75 sessions and 2.8 hours on AI apps in the first half of 2025, according to Sensor Tower, which counts Vietnam among the fastest-growing AI markets in the world.

What workers like them are experiencing now has a name. In a study published in Harvard Business Review in March 2026, a team from Boston Consulting Group surveyed 1,488 full-time U.S. workers and defined "AI brain fry" as mental fatigue from using or overseeing AI tools beyond a person's cognitive capacity.

About 14% of AI users reported it, with marketing staff hit hardest at 26%. The researchers found the strain comes less from using AI than from policing it: workers with heavy oversight demands expended 14% more mental effort and reported 19% more information overload than those whose AI work needed little supervision.

Productivity had a ceiling too, holding up to about three tools running at once and then reversing. One senior engineering manager said that switching between tools and double-checking every output left his thinking "cluttered" rather than faster.

"The paradox is that AI was created to reduce the load, but in practice it is adding pressure through a different mechanism," said Cao Tran Thanh Trung, CEO of the Lumos Psychology Center, who says he is seeing more burnout cases tied to the automation wave.

"It's an unfair race," he said. "The problem isn't the technology itself, but the way businesses are using it as an excuse to raise output without adding people."

When the nervous system stays on high alert for long stretches without real recovery, he said, it loses the ability to reset, bringing trouble concentrating, slower decisions, irritability, insomnia and nights spent awake with the mind still circling work. Left unchecked, it can deepen into depression.

His advice is not to drop the technology but to redraw the lines around it: let AI take the routine, and protect genuinely tech-free breaks so an overtaxed nervous system can recover.

Minh Thu eventually told her manager where her limits were and proposed splitting up the workflow differently. She has also stopped reaching for AI by reflex, learning when to switch it off "to protect my own capacity for original thinking."