TikTok's July 1 cuts in Singapore, Indonesia, Malaysia and Dublin show what AI restructuring really looks like: fewer human moderation jobs, even while the company keeps hiring for more technical trust and safety roles.
TikTok employees in Singapore were told on July 1 that their roles were being cut, and the company confirmed the affected jobs sat inside its Trust and Safety division. Mothership reported that one affected staffer said about 20 people on their team alone were let go. TikTok did not give a total number for Singapore, which is exactly the kind of omission that makes a restructuring feel colder for the people inside it.
This was not a one-city clean-up. TikTok also confirmed layoffs in Indonesia the same day, describing them as a realignment of its research and development organization after the company's integration with Tokopedia. Former employees in Malaysia posted about their own retrenchments. In Ireland, local reports said around 300 roles could be affected at TikTok's Dublin operation, where the company has built one of its major European trust and safety hubs.
You can see the pattern without needing TikTok to spell it out. The work being squeezed is the work that used to need large numbers of people reviewing content, training systems, checking quality and handling local queues. The work being protected, and in some cases expanded, is more specialised. TikTok told Irish media it was exploring a reorganization to strengthen its global operating model for Trust and Safety, including hundreds of new specialist roles in Dublin and possible redeployment opportunities.
That is not a contradiction. It is the trade.
TikTok has said automated systems now identify and remove more than 85% of content taken down for violating its community guidelines. That figure is the spine of the story. Once a platform can say machines already handle most removals, the old argument for keeping large distributed moderation teams gets weaker inside the finance department, even if it still makes sense to the people who understand local language, slang, politics and context.
Singapore, Jakarta, Kuala Lumpur and Dublin were not random names on a spreadsheet. These offices helped TikTok handle regional users, regional laws and regional risk. If you run a platform where one viral video can become a political problem before lunch, local context is not decoration. It is part of the product. Centralising that work into fewer hubs may be cheaper and easier to manage, but TikTok has not shown that it will be better for the countries losing reviewers.
The company is not alone, but that does not make the move harmless. The Guardian reported last year that TikTok had already put hundreds of UK moderator jobs at risk as it relied more heavily on AI, while the Wall Street Journal noted the same shift toward automated moderation. Meta has also pulled back from parts of its old content policing model, most visibly by ending third-party fact-checking in the United States. X went further after Elon Musk bought the company, cutting deep into trust and safety teams. This is where the industry is going.
Frankly, the Dublin detail is the one to watch. If TikTok can cut hundreds of roles while creating hundreds of specialist jobs in the same city, the company is not simply shrinking trust and safety. It is changing who gets to belong in it. The manual reviewer loses ground. The engineer, policy specialist and AI operations worker gain it.
For employees, that is a brutal distinction. TikTok's standard promise of support, local legal compliance and transition resources will not tell a laid-off analyst in Singapore whether their language skills are still valued, or whether the job they trained for is disappearing everywhere at once. It also will not answer the harder platform question: who catches the mistakes when automated systems move faster than human judgment?
The answer may come first from regulators, not TikTok. Europe already has the Digital Services Act. The UK has the Online Safety Act. Ireland matters because TikTok's Dublin base sits inside that regulatory map, and any weakening of moderation capacity there will be watched closely. If the company wants fewer humans doing the first pass, it will need to prove the machines are not just faster and cheaper, but accurate enough when the content is messy, local and politically sensitive.
That is the real story behind the July 1 layoffs. AI is not arriving as a future idea in trust and safety. It is already deciding which jobs remain, which jobs move, and which workers open an email in the morning to learn their role has vanished.
Also read: Kioxia Sells Out Its Entire 2026 NAND Flash Supply Just as AI Chips Ship • A Nuclear Reactor in Utah Just Powered an Nvidia Chip for the First Time • Anthropic Opens Talks With Samsung to Build Its First Custom AI Chip
