Not our race to win, but ours to lose

We finally heard the smartest technology policy out of the government in a decade, right here in Iloilo City.

During the 2026 AI Fest this week, Science Secretary Renato Solidum Jr. said the quiet part out loud: “We are not trying to out-build Silicon Valley… That is not our race to win.”

He’s right. The global AI arms race is a multi-billion-dollar bloodsport. When foreign companies pull in funding rounds topping USD 6,6 billion just to train massive frontier language models, the Philippines trying to compete head-to-head is financially absurd. Instead, DOST is making a pragmatic pivot. They want to build AI for our actual realities — technologies that recognize jeepneys instead of self-driving Teslas, and weather models trained for tropical cyclones, not temperate drizzles.

But a good keynote is just a keynote. The developer’s reality check is much harsher.

Building “AI for Filipinos” sounds deeply patriotic until you look under the hood. Models are only as good as the data feeding them. To build a vision-based traffic system like vBANT.AI that actually maps tricycles, you need millions of cleanly labeled data points from Philippine roads. Building natural-language tools like iTanong is a structural nightmare in a country that code-switches in almost every sentence across dozens of regional languages. Those of us learning to integrate these technical frameworks into our own daily workflows know firsthand how quickly foreign models break down when fed local context.

This is why consolidating our scattered government projects under the National Artificial Intelligence Center for Research and Innovation (NAICRI) isn’t just bureaucratic reshuffling. It is a desperate, necessary technical prerequisite. You simply cannot build Filipino AI on borrowed, unstructured data.

Yet, if there is a blueprint for making this work, it is quietly being written right in our backyard. Even before hosting the festival, Western Visayas through its Regional Development Council was the first to adopt both an AI action plan and a governance and ethics policy ahead of national mandates.

By putting heavy compute power directly into local institutions — think WVSU and ISAT U — we are decentralizing tech talent. We are giving local coders and researchers the iron they need to build localized tools without forcing them to migrate to Imperial Manila.

There is, however, a glaring blind spot in this roadmap.

Solidum noted that while 80 percent of Philippine establishments have basic digital infrastructure, only about 15 percent use advanced AI tools. Unsurprisingly, that uptake is hoarded by large, urban corporations.

Flagship public service tools like Project GABAY are excellent. But the private sector adoption gap means our small and medium enterprises are bleeding out. The severe shortage of specialized AI talent disproportionately hurts smaller businesses that simply cannot match corporate salaries. If the National AI Strategy (NAIS-PH) does not explicitly democratize access for the 85 percent of businesses currently ignoring this tech, our localized AI will only deepen the country’s digital divide.

AI for the Filipino should not just mean AI for the mega-corporation. Progress is about building what we actually need — but it only matters if everyone gets to use it.