Manila’s Extreme Poverty Could Be Addressed by Predictive AI Social Services Targeting
By Joel Comm | A Trusted Voice in a Noisy Tech World
"We have 4 million people living in informal settlements across Metro Manila, and our current social services reach maybe 20% of those who need help most," says Dr. Maria Santos, Director of Policy Research at the Ateneo School of Government. "The question isn't whether we have enough resources. It's whether we're smart enough about how we deploy them."
That quote hit me hard during my recent visit to Manila. I've spent 45 years watching technology promise to solve big problems, but this feels different. AI isn't just another shiny tool here. It's potentially the key to untangling one of the world's most complex poverty puzzles.
The Inequality Algorithm Is Already Running
I explore a different angle on this in my companion article, Philippines AI Summit 2026: Why This Nation's Capital Is Emerging as Southeast Asia's AI Voice.
Walk through Makati's gleaming Ayala Triangle, then take a 15-minute jeepney ride to Tondo. The contrast isn't just visual. It's algorithmic. Manila's extreme inequality operates like a machine, sorting people into haves and have-nots with mathematical precision.
Traditional social service delivery makes this worse. If you live in a barangay with good internet and educated community leaders, you'll probably know about available programs. If you're in an area where the barangay captain doesn't speak English well or lacks digital literacy, those same programs might as well not exist.
I watched this firsthand at the Department of Social Welfare and Development's main office. Families travel hours from outlying areas, often spending money they don't have on transportation, only to discover they're missing one document or don't qualify for a program they heard about through word of mouth.
The current system punishes the poorest for being poor. AI could flip that equation.
Predictive Targeting Gets Personal
Here's where it gets interesting. The Philippine Statistics Authority has been quietly building one of Asia's most comprehensive datasets on household welfare. Combined with mobile money transaction data from GCash (which serves 90% of Filipino adults), satellite imagery showing informal settlement density, and even social media activity patterns, you've got the raw materials for genuine predictive social services.
Think Netflix recommendations, but for poverty intervention.
Imagine an AI system that identifies a family in Quezon City's Commonwealth area three months before their breadwinner loses construction work due to seasonal patterns. Or one that spots households at risk of medical bankruptcy because of emergency room visit patterns combined with GCash transaction histories showing declining savings.
Smart Communications and Globe Telecom already have anonymized mobility data showing how people move through the city. Cross-reference that with economic activity indicators, and you can predict which communities will be hit hardest by the next typhoon or economic downturn.
This isn't science fiction. Similar systems are running pilot programs in Kenya and Brazil right now.
The Jakarta Problem
But here's what keeps me up at night about this technology. AI systems learn from historical data, and Manila's historical data is soaked in inequality.
If past social service delivery favored English-speaking families with formal addresses and bank accounts, an AI system trained on that data will systematically exclude Tagalog-speaking families living in areas without street names who rely on cash transactions.
I call this the Jakarta Problem, after Indonesia's disastrous attempt to automate benefit distribution. Their AI system got really good at serving people who were already being served well. Everyone else fell further through the cracks.
Manila can't afford to make that mistake.
Building Fair Algorithms From Scratch
The solution requires rebuilding social service targeting from the ground up with equity as a core design principle.
First, data collection has to be community-driven. Instead of waiting for people to show up at government offices, deploy teams with tablets to informal settlements. Train local leaders to input data in their own languages. Use image recognition to automatically process handwritten forms.
Second, the algorithms themselves need regular bias audits. If your AI system recommends more job training programs in Makati than in Tondo, that's a bug, not a feature. The goal is distributional justice, not efficiency optimization.
Third, make the system transparent to the communities it serves. Families should understand why they were selected for specific programs and have clear paths to appeal decisions.
In my other article about Manila, I explore how AI is already transforming Manila's business landscape through automation and data analytics. Those same technological capabilities could revolutionize social services if deployed thoughtfully.
Making It Work in the Real Manila
The most promising developments are happening at the barangay level. Several communities in Pasig and Marikina are piloting AI-enhanced health monitoring systems that predict which residents are most likely to need medical intervention based on environmental factors, age, and reported symptoms.
These hyperlocal systems work because they're built by people who understand their communities intimately. The AI amplifies human knowledge rather than replacing it.
Success will require unprecedented cooperation between national government agencies, local officials, technology companies, and community organizations. The Department of Information and Communications Technology would need to coordinate with DSWD, the Philippine Health Insurance Corporation, and literally thousands of barangays.
That's a massive coordination challenge. But the alternative is watching inequality compound year after year while we have the tools to intervene sitting unused.
If your organization in Manila is navigating these changes, having an ai keynote speaker in Manila who understands both the technology and the human side can make the difference. The stakes are too high for purely technocratic solutions.
AI won't solve Manila's poverty overnight. But used correctly, it could ensure that help reaches families before they fall into crisis rather than after. That's not just smarter government. It's more human government.
Get Joel’s Free Guide
AI Made Simple for Your Organization — six plain-language principles every leader needs right now. No tech lingo. No hype. Just what actually works.
