Hoy, kumusta? Welcome back to Kaisipan!
Three AI bills clash in Congress while schools rush to adopt AI without first fixing the basics.
Here’s what happened in AI this week:
Microsoft pushes AI tools into 1,500 schools.
TESDA launches all-in-one skills passport app.
Scammers weaponize AI for romance fraud.
ChatGPT embeds ads, sparking industry war.
Nvidia delays RTX 60 GPUs to 2028.
Breaking Down the 3 AI Bills in Congress and What This Means for an AI-Powered Philippines
Congress has three AI bills in right now. Each one promises to regulate, innovate, or protect Filipinos from AI harms. Pero, The issue is that each bill overlaps with the others, causing some conflicts. Here’s my breakdown.
House Bill No. 1484, the Data-Driven Governance and AI Solutions Act, focuses on government AI adoption. It creates a ₱250M fund under the Department of Science and Technology (DOST) for pilot projects like traffic systems, disaster mapping, and fraud detection. Think modernization, not regulation. This bill aims to improve public service delivery and build government capacity through regional AI governance hubs.
Senate Bill No. 25, the Artificial Intelligence Regulation Act (AIRA), is the strictest. It requires every AI system to register before being imported, sold, or deployed. No registration, no access. It creates a National AI Commission with quasi-judicial powers, sets risk tiers, and bans deepfakes, mass manipulation, and unlawful surveillance. Strong consumer protection potential, but heavy compliance burden.
House Bill No. 3195, the Artificial Intelligence Development and Regulation Act, tries to do everything: an AI Bill of Rights, multiple oversight bodies (i.e., Philippine Council on AI, AI Board, National Center for AI Research) and bans on lethal autonomous weapons and social scoring systems. It's the most rights-focused of the three but creates institutional overload.
The conflict is obvious. AIRA puts the AI Commission under DOST. HB 3195 puts three bodies under the Department of Information and Communications Technology (DICT). HB 1484 creates another steering committee under DOST. Kung pagsasabayin mo, tatlong ahensya ang mag-aaway kung sino ang may hawak ng AI policy. Both AIRA and HB 3195 require registries—dalawang databases, dalawang compliance headaches. Classic Filipino bureaucracy that slows downs process.
The problem with these bills:
AIRA's "register all AI" requirement is impossibly broad. That's not risk-based regulation; that's bureaucratic overload.
HB 3195's AI Board requires unanimous decisions from six agencies, including national security and privacy regulators. This is a recipe for deadlock.
HB 1484 only governs government-funded projects, so private sector AI harms remain unaddressed.
Here’s what's happening internationally:
The European Union AI Act regulates high-risk systems—hiring, credit, essential services, and remote biometric identification—with heavy documentation requirements.
Singapore's Model AI Governance Framework gives companies guidance on building internal governance structures, mapping AI risks, testing for bias, and maintaining human oversight.
The United States uses a voluntary AI Risk Management Framework that breaks down risk management into four functions: govern, map, measure, and manage.
The UK assigns primary responsibility to existing sector regulators—financial regulators oversee banking AI, health regulators oversee medical AI—using shared principles like safety and accountability.
The Philippines? We're building three overlapping bureaucracies before defining what "high-risk" even means for Filipinos. Generic "register everything" requirements won't catch criminals; they burden people and businesses and ultimately slow everything down.
So what needs to change?
Pick one lead regulator. Creating multiple agencies, commissions, or what is chaotic. There are too many cooks in the kitchen!
Replace "register everything" with high-risk-first compliance. Make algorithmic impact assessments mandatory with public summaries. Daily calories ko nga hirap akong i-track. Eto pa kayang register everything? Good luck.
Streamline AI regulation. Bakit ang daming bills about AI? I know each have their own purpose pero mas maganda kung isahin na lang.
Right now, these bills promise protection but deliver confusion. Fix the turf wars first. Otherwise, enforcement becomes selective, and selective enforcement in the Philippines means corruption risk.
Bes, Ano Yung…AI Governance?
AI governance is the system of rules, institutions, and accountability mechanisms that determine how artificial intelligence is designed, deployed, and controlled.
Think of it as guardrails for AI power. It answers critical questions:
Who decides what AI can do?
Who's responsible when it causes harm?
How do we prevent bias, surveillance abuse, or unchecked automation?
It operates on three layers.
Policy and regulation—governments define what's legal.
Organizational governance—companies decide how they build and use AI internally through ethics boards and audits.
Technical controls—how systems are built with explainability and human oversight.
Without governance, bias scales automatically, surveillance becomes normal, and power concentrates in the hands of whoever builds the systems. For Filipinos, governance prevents us from becoming just data sources for AI built elsewhere.
Microsoft Pushes AI Reading Tools Into 1,500 Philippine Schools
Microsoft and DepEd are doubling down on AI-powered learning tools to fix the Philippines' literacy crisis.
Reading Progress, Microsoft's AI tool, already delivered results in Bais and Dumaguete (14,000 learners assessed across 61 schools) and Cabanatuan City (100% of learners in 3 districts leveled up).
The tool automates reading checks, identifies gaps, and cuts admin time for teachers. By 2026, Microsoft aims to reach 3,000 teachers across 1,500 schools through the ARAL program.
The real question: will this actually close the reading comprehension gap plaguing Philippine education? And more critically, will it narrow the divide between public and private school outcomes?
Private schools consistently produce better results. They have resources, less bureaucracy, and faster implementation. They can expand AI offerings and pass costs to parents who won't think twice about paying higher tuition.
Public schools? Ibang usapan yan. Twenty approval layers before anything happens. Limited infrastructure. Overwhelmed, if not incompetent (real talk?), teachers. If AI tools favor schools with the capacity to deploy them quickly, the gap widens rather than closes. Execution will reveal whether that's realistic or just aspirational.
Mga Ibang Ganap
TESDA launched its Skills Passport app on February 5, integrating training, jobs, and scholarships into one platform with P20 billion allocated for workforce development.
Scam Watch PH and CICC launched Unmatch PH campaign against AI-powered love scams targeting five profiles including financially stable professionals and solo parents seeking companionship.
Nvidia delayed its RTX 60 series GPUs to 2028 due to memory shortages, prioritizing AI chip production over gaming hardware and extending RTX 50 series lifespan beyond three years.
OpenAI introduced Frontier platform helping enterprises deploy AI agents across business operations, with early adopters including HP, Intuit, Oracle, State Farm, and Uber reporting significant productivity gains.
Anthropic released Claude Opus 4.6 with improved coding skills, 1M token context window, and state-of-the-art performance on agentic coding and knowledge work evaluations beating GPT-5.2.
Mga Muni-muni
Filipinos are burning out. Steven Bartlett's podcast on ambition and burnout in the digital age is resonating globally, but the issue hits different here. We're not just dealing with hustle culture. We're living in full-blown raket culture. Multiple side gigs, freelance work across time zones, always-on digital presence. And now AI tools promise to make everything faster.
Here's the uncomfortable question: kung mas mabilis na ang trabaho dahil sa AI, bakit parang mas pagod pa tayo?
AI does make things faster. Tasks that used to take hours now take minutes. Automation handles repetitive work. The narrative says this creates opportunity—reskill, upskill, stay competitive in a changing job market. And that's true.
But here's what nobody talks about: we're not using that freed-up time to rest. We're using it to take on more work. More clients. More projects. More raket.
I get it. Most of us are breadwinners. Extra time means extra income. Extra income means helping family, paying bills, and building a financial cushion. Rest feels like a luxury we can't afford. Kapag may idle time, guilt kicks in. "Pwede pa akong kumita dito."
Did AI actually make life better? Or did it just give us more capacity to exhaust ourselves Swerte mo kung yung freed-up time mo, ginagamit mo para mag-rest or mag-improve ng sarili. But that's privilege. The rest of us? We're just grinding harder because the tools let us. Technology promised freedom. But without intentional boundaries, it just became another way to work more.
Prompt Tip!
Everyone's talking about AI agents, but most explanations are either too basic or too technical. The truth? You probably don't need an agent yet.
Here's the framework from productivity expert Enovair:
Start with level 1 automation (if this, then that).
Move to level 2 AI workflows when you need reading and categorization.
Only build level 3 agents when you need autonomous decision-making with unpredictable inputs.
Most businesses can solve 80% of tasks with simple workflows. Look for automation signals first: repetitive tasks, predictable outcomes, clear start and end points. Map one process, identify manual steps, then pick the simplest tool that works.
The key insight? Workflow first, agent light. Don't overcomplicate. Build custom only when your existing tools can't handle it. Start strategically with low-stakes tasks and proper guardrails.
That’s all for today!
Thank you for reading today’s newsletter!
Do you have suggestions? I’d love to hear your thoughts.