Philippines proposes 6 months' notice before AI-driven layoffs: jobs are protected with time, not brakes

By BigGo Finance · November 8, 2026.

Philippine Senator Erwin Tulfo has filed Senate Bill No. 2563, which would amend Article 298 of the country's Labor Code to regulate layoffs linked to artificial intelligence. According to the text as described by BigGo Finance, employers would have to notify the Department of Labor and Employment (DOLE) at least six months before AI-driven job cuts.

The proposal adds further requirements. Consultations with affected workers would be mandatory. Retraining and redeployment would be prioritized. A dismissal would only be valid if the AI adoption is genuine, in good faith and backed by a real economic justification, such as cost savings or efficiency, and if no other alternative exists. In addition, selection criteria would have to be fair and reasonable.

Those who lose their jobs would receive severance pay of at least two months' salary, or two months per year of service, whichever is greater. Tulfo, who chairs the committee on social justice, welfare and rural development, justifies the measure in his explanatory note: technological progress is inevitable, but Filipino workers should not bear its cost alone.

The bill still has to pass the Senate and the House of Representatives, and its prospects are uncertain. According to the source, employer groups will likely push back on compliance costs and hiring flexibility. Malacañang, the presidential palace, has acknowledged that AI may displace workers but also open up opportunities.

Context matters. The Philippines relies heavily on business process outsourcing (BPO) and back-office services, the kind of administrative work that our analysis of employment considers most exposed to automation. The source itself notes that the law would add a layer of procedural protection for workers in these sectors, which are heavy users of automation.

Our take: the bill is a good sign but has a design flaw. The good sign is that governments are starting to treat the transition as a management problem rather than an abstract debate. Six months' notice, consultation and retraining honestly acknowledge that the short term will be hard and that someone has to cushion the blow. That is preferable both to ignoring the problem and to banning the technology.

The flaw is the "last resort" clause. Requiring a company to prove that no alternative exists is a test that is hard to verify and easy to litigate. It could turn automation into a legal ordeal and push companies to restructure abroad or avoid hiring in the first place. If the goal is to protect the person rather than the specific job, notice, severance and genuine retraining matter more than a de facto veto.

It is also worth separating what the text promises from what it will achieve. Mandatory retraining only works if there is somewhere to redeploy people: training a call center agent for a role that won't exist in two years either is meaningless. Success here will depend on DOLE and companies linking training to real demand.

If the bill passes, the winners are back-office workers, who gain time and a cushion. The loser is the company planning a quick cut, and in the long run perhaps the country itself, if the law makes AI investment too expensive compared with neighboring competitors.

The underlying thesis is ours: the transition will be difficult in the short term and warrants cushions, but the abundance AI promises in the long term is reached by adopting the technology, not by holding it back. Good regulation is regulation that shares the cost of the transition without blocking it. This bill points in that direction, although its current wording risks falling short.