āļø Authorās Note
Since 2023, Iāve spent countless hours working with AI, more, in fact, than Iāve spent with most humans. In that time, Iāve learned to prompt, debug, and collaborate with these systems in ways that feel second nature. But Iāve also encountered resistanceāfriends, colleagues, even strangersāwho question my enthusiasm. Some fear AI. Others dismiss it. And often, I find myself disagreeing with people who havenāt explored it as deeply, yet speak with certainty about its dangers.
This article is for them.
Not to prove them wrongābut to say: I understand your skepticism. Iāve written about it before in āAI Skepticism in the Philippinesā, where I explored why many Filipinos hesitate to trust what they donāt fully understand. That hesitation is valid. But so is curiosity. And empathyāon both sidesāmight be the bridge.
Why stepping into someone elseās perspective isnāt just good for peopleāitās essential for AI, too
Two students argue. One sees a white ball. The other sees a black one. Theyāre both sure theyāre rightāuntil a teacher makes them switch places. Only then do they realize: they were both correct, just from different angles.
This viral lesson in empathy, dramatized in āA Wise Lesson in Empathyā, might seem like a simple classroom skitābut it holds surprising lessons for how we build, interpret, and apply artificial intelligence in the Philippines.
Because the truth is: AI sees the world from one side of the ballāthe side itās trained to understand. If we donāt switch perspectives, we might keep arguing with it, misunderstanding it, or even fearing it.
š§ The Filipino Dilemma: AI Misunderstands Our Context
- Ever asked an AI to generate a tricycleāand it gave you a motorbike with a box?
- Or prompted it to translate a Filipino idiomāonly to get a gibberish response?
- Or tried to explain ādiskarteā and the model gives you a list of driving techniques?
These are empathy failures. Not emotional onesābut data-based ones. The AI hasnāt been taught to āseeā from the Filipino side of the ball.
If we expect AI to serve our needs, our culture, our language, it needs to be trained with our perspectives. Otherwise, itāll keep answering in ways that feel distant, alien, or worseādangerous.
š AI Alignment = Digital Empathy
What we call āAI alignmentā in technical circles is really a form of encoded empathy. Itās asking the machine:
- Do you understand what I value?
- Do you see what Iām trying to do?
- Can you adjust your behavior based on my context?
Without this, even powerful AI becomes frustratingāor even harmful. Like a helpful assistant who keeps offering soup when youāve asked for rice.
šµš So What Can Filipinos Do?
- Contribute to Local Datasets. We need open-source Filipino dataālanguage, culture, medicine, governance, diskarteāto train models that understand our reality.
- Use Prompting as Perspective-Shifting: Learn to tweak prompts like youāre switching sides of the ball. Donāt just say āsummarize thisāāsay āsummarize this for a Filipino student in Grade 10.ā
- Teach AI How We Think. Through usage, we influence outputs. Every clarification we offer (āwhen I say jeep, I donāt mean off-roaderāI mean public transportā) fine-tunes the systemās understanding of our world.
- Advocate for Culturally Aligned AI. Push institutions to develop LLMs and copilots that reflect our multilingual, multi-faith, deeply relational society. Taglish shouldnāt be a bugāit should be a feature.
š„ Final Thought: Whose Side of the Ball Does AI See?
Empathy isnāt just a feel-good lesson. Itās a survival traitāfor humans and for the digital systems weāre embedding into our banks, barangays, and classrooms.
If we want AI that sees things from our side of the ball, we must do more than consume itāwe must teach it. Shape it. Guide it with the same calm wisdom that the teacher used in that 2-minute video.
Because in the age of intelligenceāartificial or otherwiseāunderstanding isnāt about choosing sides. Itās about learning how to switch them.
