Garbage in, garbage out
Their intelligence is borrowed, their confidence is synthetic. They learn from us, our brilliance, our biases, and our bottomless capacity for nonsense.
Text size
Their intelligence is borrowed, their confidence is synthetic. They learn from us, our brilliance, our biases, and our bottomless capacity for nonsense.
Three decades after the Marcopper mining disaster devastated Marinduque’s waterways, the province is again pressing the…
Housing is probably the biggest issue affecting people the world over.
Malacañang on Monday called for a more thorough investigation into the questioned flood control projects in Taguig…
The defense on Monday backed the move by senator-judges to exclude a prosecution witness who testified on the firearms…
BORACAY — Herns Hermida of Cisco said the quiet part out loud: Artificial intelligence has already eaten the internet. Every post, review, article, and drunken rant uploaded by humankind has been scraped, sorted, and digested by machines now pretending to be smarter than their creators.
Imagine that: the sum of human chatter, preserved for eternity and regurgitated as algorithmic wisdom. And yet, Hermida added, AI still isn’t mature. Not the technology, not the people using it.
At the 2025 CXO Tech Summit in Shangri-La Boracay, Hermida joined Fortinet’s Kelvin Chua and VST-ECS president Jimmy Go in repeating the industry’s favorite confession: garbage in, garbage out.
That’s the oldest rule in computing, and somehow the most relevant again, especially in an age drowning in garbage.
Wearing my tech editor’s hat, I asked if AI had learned enough to tell the difference between trash and treasure. After all the training, fine-tuning, and alignment, could it finally filter the nonsense from the nuance?
Hermida didn’t hesitate: not yet. He broke it down simply. First, is the tech good enough? Second, are people wise enough to use it? On both counts, the answer is still “no.”
Nearly a century since Alan Turing, whom the Cisco executive cited, dreamed up the thinking machine, we are still teaching our creations how to think and quietly wondering if we are still the smarter ones.
AI tools like ChatGPT, Perplexity, and Copilot now write, summarize, and moralize, but they remain reflections of their teachers and users. Their intelligence is borrowed, their confidence synthetic. They learn from us, our brilliance, our biases, and our bottomless capacity for nonsense.
Hermida noted that these systems have absorbed nearly everything the internet has to offer, and still they hallucinate. That is the industry term for lying, though “digital bluffing” might be more honest.
On the adoption side, Hermida explained that many companies are still experimenting, caught between curiosity and confusion.
Schools even banned ChatGPT when it first appeared, afraid it would replace critical thinking. Singapore, meanwhile, took the opposite view: use it or get left behind. Fear, after all, has never stopped progress; it only delays it.
Then there was Jimmy Go of VST-ECS, whose company distributes technology like others distribute bread. Hosting CXO for the eighth year, Go offered a practical take. AI, he said, helps his teams write, check contracts, create videos, and even generate voices. Efficiency at scale.
But Go added a warning: Without knowing your own data, you’re just automating confusion. Translation: AI might speed things up, but if your inputs are garbage, you’re just producing polished trash.
Efficiency, he said, should reduce costs. Fair enough. But efficiency also has a habit of trimming more than budgets; sometimes it trims people. Oh well, maybe not at VST-ECS which seems to be perpetually on the hunt for the best people and best products and services to distribute.
On the security side, Fortinet’s Kelvin Chua brought 35 years of battle scars. His firm has used AI for two decades, not to make content or chat, but to detect the unknown.
“It’s impossible for humans alone,” he said, describing the avalanche of cyber threats launched daily. The irony is that hackers now use AI too. The same tools that defend networks are also breaking into them.
Still, Chua insisted AI would not replace people, only enhance them. A comforting thought — though that “provided they know what they’re doing” qualifier does a lot of the heavy lifting.
VMware’s Aris Coronel added the philosophical coda. AI has outgrown its toddler phase. It no longer needs more toys; it needs supervision. Governance, not to slow innovation, but to make sure the creators stay accountable for their creations.
By the end of the session, the takeaway was brutally simple. AI is not an experiment anymore. It is here, powered by everything we have ever typed, uploaded, or overshared. And it is only as good, or as bad, as the data it eats.
Garbage in, garbage out. The phrase was once about machines. Now it is about us, our attention, our discourse, and our values. We built a system that learns from every word and data we feed it.
