Laban, Pilipinas! How Filipinos Can Survive — and Win — in the Next Two Years of AI & Superintelligence (2026–2028)

Laban, Pilipinas! How Filipinos Can Survive — and Win — in the Next Two Years of AI & Superintelligence (2026–2028)

A Practical Survival Guide for BPO Workers, OFWs, Freelancers, Students, Teachers, Nurses and Filipino Families

Filipino workers have reinvented themselves before — from farms to factories, from offices to headsets, from Manila to the world. AI is the next reinvention, not the end of the story. Illustration for this paper.

Prepared October 2026. All statistics are cited to ILO, IMF, WEF, PSA, BSP and IBPAP sources listed at the end. Forecasts are labelled as forecasts. This paper gives general information, not financial, medical or immigration advice, and no buy or sell recommendation.

📄 Download the full paper (PDF, 13 pages, with charts)

How to use this paper — and what is inside

You do not need to read this cover to cover. Find yourself in the table below, read Chapter 1, then jump to your playbook in Chapter 9. Come back for the chapters in between when you have time — they explain why the playbook says what it says.

Contents

• The honest headline — transformation, not apocalypse
• The next two years: agents now, superintelligence later?
• Why the Philippines is so exposed
• Jobs hollowing out vs. jobs growing
• BPO deep dive: the ceiling just got lower
• OFWs and freelancers: the world is repricing routine work
• Money: earning with AI, and learning for free
• Risks beyond jobs: scams, deepfakes, elections, privacy
• The 2-Year Playbook: 90 days, 6, 12 and 24 months
• Honest uncertainties — and sources

A note on language. This paper is in accessible English with light Taglish, the way many Filipinos actually talk about work. Technical terms are explained the first time they appear. AI means artificial intelligence. GenAI means generative AI — systems like chatbots and image generators that create text, audio, images and code. An AI agent is software that does a multi-step task on its own, not just answers one question. AGI means artificial general intelligence, a system as capable as a human across most intellectual work. Superintelligence would go beyond that. We are not there yet — Chapter 2 shows how far apart the forecasts are.

1. The honest headline: transformation, not apocalypse

Exposure is not the same as disappearance. That single distinction, made carefully by both the International Labour Organization (ILO) and the International Monetary Fund (IMF), is the most important idea in this paper.

The ILO estimates that 28.1% of Philippine employment — about 12.7 million workers — is in occupations with more than minimal exposure to generative AI. That is the second-highest share in ASEAN, after Singapore. But only 3.3-3.6% of jobs fall in the highest-exposure tier, where displacement risk is elevated. The vast majority face partial task automation: some tasks speed up or disappear, the job itself changes. [1]

The IMF, using a different method, found 36% of Philippine jobs highly exposed. Of the whole workforce, about 22% are high-exposure and high-complementarity — AI is likely to augment them and raise their productivity — while about 14% are high-exposure and low-complementarity, the group genuinely susceptible to displacement. In plain language: for every Filipino worker in the danger zone, there is more than one whose job AI is more likely to upgrade than erase — if they learn to use it. [2]

And a reality check against panic: when Philippine unemployment rose to 5.3% (2.77 million people) in August 2026, with 49.79 million employed, the Philippine Statistics Authority (PSA) attributed the rise to absorption and manufacturing factors — not to AI. As of the latest official data, there is no evidence of mass AI job losses in the Philippines. Employment in highly exposed occupations has, so far, kept growing across ASEAN. [1][3]

That is not a reason to relax. It is a reason to use the window. The ILO’s warning is about direction: entry-level rungs are thinning (Chapter 4), women are exposed at roughly double the rate of men because they are concentrated in clerical, administrative and service roles, and regions built on digital services — the National Capital Region, Central Luzon and CALABARZON — feel the change first. Preparation during a transformation is cheap. Preparation after a displacement is not.

The Filipino advantage is real, too. English proficiency, service warmth (malasakit), adaptability, a huge global diaspora, and two decades of BPO experience in serving foreign clients are not things a chatbot copies overnight. The goal of the next two years is to pair those human strengths with AI fluency — to become the worker, freelancer or business that supervises the AI, handles the exception, owns the client relationship and is accountable for the result.

2. The next two years: agents now, superintelligence later?

Forget robots walking around. The change arriving in 2026-2028 is quieter and faster: AI agents embedded in the software you already use — answering tickets, reconciling invoices, drafting lesson plans, triaging chats, writing first-pass code, summarising calls — with a human approving, correcting or escalating. One agent is a helper. A company running hundreds of agents needs fewer people doing routine processing, and more people designing, checking and handling what the agents cannot.

Source: Public forecasts as stated by each forecaster. These are opinions and prediction-market medians, not facts. The disagreement itself is the finding.

How to read this table: the people building the technology disagree with each other by six years or more. Anyone who tells you a certain date — in either direction — is selling certainty nobody has. Plan in layers instead:

• Certain enough to act on now: agents will absorb more routine digital tasks in 2026-2028; employers will expect AI use; entry-level hiring will demand more from day one.
• Plausible, prepare lightly: whole workflows (not just tasks) get automated in customer support, back office, content and basic software work.
• Uncertain, do not bet your life on it: AGI or superintelligence arriving inside this window and making all human work obsolete. If it happens, no individual career plan survives it anyway — government and global policy would matter more than your resume. Your best hedge is the same as for the likely case: savings, adaptable skills, and human work AI complements.

3. Why the Philippines is so exposed

The Philippines is exposed for the same reasons it prospered: it sells services, English and trust to the world — exactly the work generative AI touches first. Three pillars carry the economy, and AI presses on all three.

Source: ILO ASEAN GenAI research, July 2026.

Source: IMF Working Paper WP/25/43, February 2025. Shares of the total Philippine workforce.

Pillar 1 — IT-BPM / BPO. About 1.9 million jobs and over $40 billion in revenue in 2025, roughly 8% of GDP (IBPAP). The IMF identifies BPO as the sector with the highest proportion of jobs at risk: about 89% of BPO roles are highly exposed in its mapping, and BPO workers — roughly 3% of the national workforce — are concentrated in the high-exposure, low-complementarity group. [2][4]

Pillar 2 — OFW remittances. Cash remittances hit a record $35.634 billion in 2025, up 3.3%, equal to 7.3% of GDP (BSP), from roughly 2.05-2.19 million overseas Filipino workers (PSA) — including seafarers, about 28% of the world’s total. AI does not sail a ship or nurse a patient, but it is changing hiring, documentation, dispatch, and the white-collar OFW jobs in the Gulf, Singapore and North America that many families depend on. [5][6]

Pillar 3 — A young, services-heavy workforce. Services account for the majority of Philippine employment, and the World Economic Forum (WEF) expects 68% of Filipino workers to need training for evolving skills — above the 59% global figure — while only 38% have completed training so far, below the 50% global figure. The intent is there: 96% of Philippine employers in the WEF survey prioritise reskilling current staff, above the global average. The gap is execution. [7]

The gender gap deserves naming. The IMF found roughly 50% of women’s jobs highly exposed, versus about 25% of men’s, and the ILO finds women at about double men’s rate in the highest-risk tier (5.85% vs 2.15% in its GenAI index). This is because clerical, administrative, sales and service jobs, where Filipinas are over-represented, are the most automatable on paper. Any serious family plan puts daughters’ and mothers’ upskilling first, not last. [1][2]

4. Jobs hollowing out vs. jobs growing

The pattern worldwide is a squeeze in the middle of routine work — and a specific, painful squeeze at the bottom rung. Stanford’s “Canaries” research on US payroll data found workers aged 22-25 in the most AI-exposed jobs saw relative employment fall about 13%, and up to roughly 19% on some measures, while experienced workers in the same jobs stayed flat or grew. Employers are hiring fewer beginners, because AI now does beginner tasks — the number to watch for a country with hundreds of thousands of fresh graduates a year. [8]

Hollowing out

Growing

Source: WEF Future of Jobs Report 2025. Global figures — new jobs do not automatically appear in the same country, city or skill level as displaced ones. The WEF also expects 39% of core job skills to change by 2030, and 41% of employers plan workforce reductions where AI automates tasks.

5. BPO deep dive: the ceiling just got lower

The BPO floor is not emptying — it is changing what it pays for: fewer headset hours, more exception-handling, domain knowledge and accountability. Illustration for this paper.

Sources: IBPAP 2025 results and July 2026 revised outlook.

Read those two charts together, because they tell the true story. The industry is still growing — from about 1.9 million jobs and over $40 billion in 2025. But in July 2026, IBPAP cut its 2028 ambition: from the 2022 roadmap’s $59 billion and 2.5 million jobs down to a best case of $50.5 billion and 2.14 million jobs, with a downside of $43.3 billion and 1.85 million jobs — essentially flat headcount. [4]

Translation for an agent on the floor: your company can earn more while hiring fewer people like the person you were hired as. Growth is moving to Global Capability Centres — the in-house Philippine units of multinationals doing finance, HR, IT and engineering work — which hire for domain expertise and ownership of outcomes, not headset hours. IBPAP reports affected entry-level staff have so far mostly been redeployed rather than let go, and the industry target is a workforce that is AI-enabled across the board — which means basic “AI fluency” will soon earn no premium at all. The premium moves above fluency: healthcare accounts, banking and insurance knowledge, technical support depth, quality assurance of AI outputs, team leadership, and being the human a client trusts when the agent fails.

6. OFWs and freelancers: the world is repricing routine work

OFWs — the money is strong, the ground is shifting. A record $35.634 billion in cash remittances in 2025 (BSP) says Filipino labour abroad is still in demand. Hands-on work — nursing and care, seafaring, construction, hospitality, domestic work done with trust — has low direct GenAI exposure. The risks are indirect: AI-screened recruitment (your application is read by software first), automated documentation and dispatch, shrinking white-collar headcounts in host countries, and — for seafarers — slow but real automation in navigation support and port operations. The OFW hedge is credentials: keep licences current, add the digital skills your specific industry is adopting (electronic health records, maritime digital systems, computerised maintenance), and convert remittances into assets and skills at home rather than pure consumption, so a lost contract is a setback, not a catastrophe. [5][6]

Source: Upwork demand / postings data as reported for AI-era categories. Direction matters more than the exact month: clients pay more for AI-augmented outcomes and less for raw routine output. AI-skilled freelancers were reported to earn about 34% more per hour.

For freelancers and MSMEs, the winning offer is no longer “I will write it” — it is “I will deliver the finished, checked result, faster, using AI under my supervision.” Illustration for this paper.

Freelancers — the split is already in the data. On Upwork, postings for writing fell about 32%, translation 20% and customer service 16%, while AI video generation rose about 329%, AI integration 178% and data annotation 154%. Clients have not stopped buying; they have stopped buying unaided routine output at old prices. A Filipino VA who only schedules and transcribes competes with an agent. A VA who runs a client’s whole content pipeline, store operations or lead follow-up — using AI, checking its work, reporting results — sells an outcome, and outcomes still command a premium. [9]

7. Money: earning with AI, and learning for free

Start with what is free. TESDA aims to equip one million workers with AI skills through micro-credentials and online learning, with P641 million for AI and data training for 35,000 learners, and a joint TESDA-DICT programme of about P740 million targeting 340,000 scholars over three years. [10]

Should you pay for AI tools? Put the price in perspective, honestly. NCR minimum wage is now ₱755 per day for non-agriculture (₱718 for agriculture, small retail/service and small manufacturing) under Wage Order No. NCR-28 issued by RTWPB-NCR, effective 26 September 2026 — up from ₱695 (₱658) under Wage Order NCR-26. The higher ₱780 tranche in the separate Wage Order NCR-27 is restrained in litigation and is not in effect. ChatGPT Plus at about ₱999 per month (Philippine App Store local pricing) costs roughly a day and a half of minimum wage; Claude Pro at about ₱1,398 per month including VAT costs about two days. As a guide, $20 USD base subscriptions billed direct on the web generally fluctuate between ~₱1,150 and ~₱1,350 depending on USD/PHP FX rates, before/with 12% digital VAT depending on billing route. (Prices depend on current FX rates and app store vs. direct web VAT billing). For a freelancer whose income depends on output, one paid tool used daily can be a rational work expense. For a student or a family under pressure, it is not: start free, and pay only when you can name the feature that earns more than the fee. Never borrow for tools or “AI business” packages. Walang easy money.

8. Risks beyond jobs: scams, deepfakes, elections, privacy

For many families, the first AI harm will not be a lost job. It will be a phone call.

Five more defences that cost nothing

• Verify before you share. AI-generated election content will surge into the 2028 elections: fake endorsements, synthetic “leaks”, cloned candidate voices. Check the original source, the date, and at least one independent outlet before forwarding. If a clip makes you furious in ten seconds, that is the design — pause for ten minutes.
• Protect your voice and face. Public videos are training material for clones. Keep accounts private where you can, and teach grandparents that a familiar voice on the phone is no longer proof of identity.
• Mind what you paste. Never paste company data, client names, passwords, government IDs or family documents into a free AI chatbot. Assume anything you paste can be stored. Use employer-approved tools for employer work.
• Guard the digital divide in your own barangay. ILO exposure is highest in NCR, Central Luzon and CALABARZON — and opportunity follows connectivity. A relative in the province with a phone but no data, no laptop and no English practice faces a double gap. Sharing devices, data and skills inside the family is not charity; it is how the whole family’s risk falls.
• Teach children the difference between fluent and true. AI answers confidently and sometimes wrongly. The school skill of the next two years is verification: ask for sources, cross-check, and show your work. Teachers: the free Google AI Educator Series (Chapter 7) is built for exactly this.

Privacy, briefly. The Philippines has a Data Privacy Act and a National Privacy Commission. If an employer or client asks you to feed personal data into an AI tool, ask which tool, whether it is approved, and where the data goes — in writing. “The AI did it” will not be accepted as an excuse by a client, a regulator, or a court, and it should not be accepted in your family group chat either. The human who pressed send stays responsible. That, in the end, is also the Filipino worker’s selling point: accountability a machine cannot offer.

9. The 2-Year Playbook — start with the next 90 days

Every persona below follows the same rhythm. Do not skip the 90-day block: it is deliberately small, because a finished small plan beats an abandoned grand one.

The universal rhythm

Playbook A — BPO agent

• 90 days: Learn your company’s approved AI tools properly; volunteer for AI-pilot queues; finish one TESDA Online AI course; start the emergency fund, even at P500 per payday.
• 6 months: Pick your account’s domain (healthcare, banking, insurance, logistics, tech) and learn its vocabulary and rules; ask HR in writing about registered enterprise training and GCC openings.
• 12 months: Move toward QA, training, team-lead support or complex-case handling; document measurable gains (handle time, CSAT, accuracy) for appraisal.
• 24 months: Be internal-first for a GCC or specialist role; keep an updated resume and one outside option warm at all times.

Playbook B — Freelancer / virtual assistant

• 90 days: Rebuild one offer as an outcome (“monthly content pipeline, checked and scheduled”) instead of hourly tasks; finish DICT SPARK or a TESDA course; raise no prices yet — raise proof.
• 6 months: Add one surging skill: AI video, AI integration for small businesses, or data work; collect testimonials tied to results.
• 12 months: Productise: fixed packages, two or three clients maximum dependence on any one; set peso and dollar income targets separately.
• 24 months: Specialise in one industry you know deeply; your moat is the industry’s trust, not the tool.

The Playbook, continued

Playbook C — Student / fresh graduate

• 90 days: Use AI to learn faster, not to skip learning — draft with it, then verify and rewrite in your own words; start a portfolio folder today, even with school projects redone properly.
• 6 months: Finish one free credential (TESDA x AWS track, Google Career Certificate with financial aid, or DICT SPARK); join one real project — a barangay, church, school or small business that needs digital help.
• 12 months: Graduate with three things: your diploma, a portfolio link, and one reference who watched you deliver real work. That trio answers the entry-level squeeze directly.
• 24 months: Target growing roles (Chapter 4, right table) and GCC / specialist employers; treat your first job as paid training and keep learning one skill per quarter.

Playbook D — OFW / seafarer

• 90 days: Set the family code word and scam rules (Chapter 8) on your next call home; check that your licences, sea service / employment records and IDs are complete, digital-backed-up and not expired.
• 6 months: Add your industry’s digital layer: electronic health records for health workers, digital navigation and maintenance systems for seafarers, computerised systems for technicians; learn how AI screening reads applications — clean, keyword-honest, verifiable documents.
• 12 months: Agree with your family on a written split for every remittance: needs, emergency fund, and one investment in skills or a small asset at home. Consuming 100% of a record remittance year is the real risk.
• 24 months: Hold a “Plan B country / Plan B role” on paper; keep credentials portable so one lost contract does not end the family’s income.

Playbook E — Parent, teacher, nurse, MSME owner

• 90 days: Parents: code word + one family AI hour a week on a free course. Teachers: start the Google AI Educator Series. Nurses: master your facility’s digital records and AI-assisted tools — clinical judgment plus digital fluency is the premium. MSME owners: put your business on maps and messaging apps properly, with AI helping draft replies and listings — checked by you.
• 6 months: Parents: guide children to one portfolio project, not more screen time. Teachers: build one AI-assisted lesson workflow you can defend academically. MSMEs: automate one routine only — replies, inventory list, or bookkeeping summary — and measure if it saved time.
• 12 months: Turn the experiment into an asset: a credential, a classroom practice other teachers copy, a store with reviews and repeat online customers, a care specialisation.
• 24 months: Teach someone else. In every Filipino success story in this transition, the last step is the same: the person who learned becomes the person who teaches — at home, in school, in the barangay. That is how a family’s risk becomes a community’s advantage.

10. Honest uncertainties

A paper that asks for your trust should end by saying what it does not know.

• We do not know the AGI date. Chapter 2’s forecasts span 2026 to 2033 and beyond. This plan is built to be right under the slow scenarios and useful under the fast ones — it is not built on any single date.
• Exposure studies are maps, not weather forecasts. ILO and IMF percentages measure what AI could do to tasks. What employers actually do depends on cost, regulation, client trust and worker organising.
• So far, transformation is outrunning displacement — in PSA data, in ILO’s ASEAN finding that exposed jobs keep growing, and in IBPAP redeployment reports. “So far” is doing real work in that sentence. Watch entry-level hiring, BPO headcount and PSA releases, not social-media claims.
• Global net gains hide local pain. The WEF’s net +78 million jobs is a world total. A displaced agent in Cebu is not comforted by a new AI job abroad — which is why this paper’s unit is the Filipino family and its next 90 days.
• Prices, wages and programme details change. Tool prices, the NCR wage and programme eligibility were as reported in October 2026. Verify before you pay or enrol.
• No one can promise you income — not a course, a tool, or this guide. Preparation changes your position when change arrives, and position is most of the game.

Last word. Filipinos have crossed oceans, rebuilt after storms and sent $35.6 billion home in a single year. Learn AI, check it, teach it — and do not let anyone convince you that you are already left behind. Hindi tayo maiiwan — kung kikilos tayo ngayon. Laban, Pilipinas.

Sources and endnotes

[1] ILO — Generative AI and jobs in the Philippines; ILO ASEAN GenAI research, July 2026: 28.1% of Philippine employment with more than minimal GenAI exposure (12.7M workers), 2nd in ASEAN after Singapore (42.2%); 3.3-3.6% in highest-exposure tier. [2] IMF Working Paper WP/25/43, February 2025: 36% highly exposed; ~22% high-exposure/high-complementarity; ~14% high-exposure/low-complementarity; ~50% of women’s jobs highly exposed vs ~25% of men’s; BPO ~89% of roles highly exposed. [3] PSA Labour Force Survey, August 2026: unemployment 5.3% (2.77M), employed 49.79M; rise attributed to absorption/manufacturing, not AI. [4] IBPAP: IT-BPM 1.9M jobs, over $40B revenue, ~8% of GDP in 2025; July 2026 revised 2028 outlook — best $50.5B / 2.14M, downside $43.3B / 1.85M, vs 2022 roadmap $59B / 2.5M. [5] BSP: OFW cash remittances $35.634B in 2025, +3.3%, 7.3% of GDP. [6] PSA Survey on Overseas Filipinos: ~2.05-2.19M OFWs; seafarers ~28% of world total. [7] WEF Future of Jobs Report 2025: 170M created, 92M displaced, net +78M by 2030; 39% of core skills change; 41% of employers plan AI-related reductions. [8] Stanford Canaries: ages 22-25 in most AI-exposed US occupations, ~13% relative decline (up to ~19%). [9] Upwork: writing ~-32%, translation ~-20%, customer service ~-16%; AI video ~+329%, AI integration ~+178%, data annotation ~+154%; AI-skilled freelancers ~+34%/hour. [10] TESDA / DICT / DepEd / CHED 2025-2026: TESDA x AWS free via TESDA Online (Sept 29, 2026); DICT SPARK; Google AI Educator Series (Oct 2026). [11] Scam reporting: ~52% scammed at least once, average loss P11,896. Forecasts Ch.2: Amodei, Altman, Hassabis; Metaculus median full AGI January 2033.