The AI's thinking now becomes your thinking and you don't get a choice.
On April 27, nearly five hundred members of the University of Hawaiʻi community gathered virtually to hear Microsoft's AI Innovation Officer Michael Jabbour offer a quiet but urgent reminder: as artificial intelligence becomes as routine as email, the greatest risk is not that machines will think for us, but that we will let them. Jabbour's message was less a technical briefing than a philosophical one — that human agency, curiosity, and the discipline to think before asking are precisely what the coming era of automation cannot replace. In a moment when seventy percent of hourly work tasks may soon be automated and employers already demand AI fluency, the university finds itself not at the edge of a technology decision, but at the threshold of a deeper question about what it means to remain the author of one's own reasoning.
- The urgency is real: up to 70% of work tasks could be automated in the near term, and over 60% of employers already won't hire candidates without AI skills — leaving institutions that don't adapt producing graduates the market won't want.
- The deeper disruption is cognitive, not mechanical — when people hand questions to AI without first thinking them through, they risk outsourcing not just labor but judgment itself.
- Jabbour's 'teach me, don't tell me' framework offers a practical counter-move, urging users to engage AI as a learning partner rather than an answer machine, keeping human reasoning in the driver's seat.
- The University of Hawaiʻi is already building infrastructure for this shift, including a free twelve-chapter AI course through the UH OneAI Hub, signaling that leadership sees this as a sustained transformation, not a one-time adoption.
- The trajectory points toward a workforce defined not by technical speed but by the human capacity to ask better questions, direct AI toward meaningful goals, and stay genuinely curious — skills no automation can replicate.
When nearly five hundred University of Hawaiʻi community members logged in on April 27 to hear Microsoft's AI Innovation Officer Michael Jabbour speak, they came expecting a technology briefing. What they got was closer to a philosophy lecture — and the distinction turned out to be the whole point.
Jabbour's central argument was deceptively simple: AI can generate limitless output, but humans must remain anchored to what actually matters. He described teaching his own children to engage AI with a single guiding phrase — 'teach me, don't tell me' — as a way of preserving the habit of thinking before asking. Hand a question to a machine without first wrestling with it yourself, he warned, and you've quietly surrendered your reasoning to the algorithm.
UH President Wendy Hensel set the institutional stakes plainly: the question is no longer whether to adopt AI, but how to do so responsibly. The university has been laying groundwork for that answer, including a free twelve-chapter interactive course called Artificial Intelligence for Hawaiʻi, available through the UH OneAI Hub — a signal that leadership understands this is an ongoing transformation, not a single technology decision.
The numbers Jabbour cited were stark. Up to seventy percent of hourly work tasks could soon be automated, AI is already matching human performance in narrow domains, and more than sixty percent of employers say they won't hire without AI competency. The institutions that fail to adapt, he implied, will fail their students.
Yet when asked what students should actually do, Jabbour bypassed coding bootcamps and certifications entirely. The future, he said, belongs to people who can articulate what they want and direct AI toward it — not to those who perform routine tasks faster than a machine. His closing thought said it plainly: 'Think before you ask.' It is, in the end, the difference between using a tool and being used by one.
Nearly five hundred people logged in from across the University of Hawaiʻi system on April 27 to hear Michael J. Jabbour, Microsoft's AI Innovation Officer, talk about what happens when artificial intelligence becomes as ordinary as email. The conversation that followed was less about the technology itself and more about how humans should think while using it—a distinction that matters more than it might seem.
Jabbour's core argument was simple but unsettling: AI can produce endless output, but humans have to stay anchored to what actually matters. He described how he teaches his own children to interact with AI models by asking them to "teach me, don't tell me"—a small phrase that carries real weight. The idea is that if you hand a question to an AI without first thinking it through yourself, you've outsourced your thinking. The machine's reasoning becomes your reasoning. You've lost the chance to choose.
UH President Wendy Hensel framed the moment as one of institutional reckoning. The question facing the university, she said, isn't whether to adopt AI—that's already decided—but how to do it responsibly. The system has been building infrastructure for this conversation since last year, including a free twelve-chapter interactive course called Artificial Intelligence for Hawaiʻi, available through the UH OneAI Hub. It's the kind of groundwork that suggests the university understands this isn't a one-time technology adoption but an ongoing shift in how people work and learn.
Jabbour offered a specific warning about the pace of change. Within the near term, as much as seventy percent of the work done in any given hour could be automated. Models are already matching human intelligence in narrow, specific tasks. At the same time, more than sixty percent of employers have signaled they won't hire people without AI skills. The math is stark: institutions that don't adapt their training will produce graduates unprepared for the actual job market.
When asked how students should prepare, Jabbour didn't point to technical certifications or coding bootcamps. Instead, he emphasized adaptability, curiosity, and human connection. The future of work, he suggested, will reward people who can articulate what they want and direct AI toward those goals—not people who can do routine tasks faster than a machine. It's a reframing that puts the burden back on humans to stay engaged, to ask better questions, to think before they speak.
Jabbour's closing thought captured the tension: "Think before you ask." Whether you're talking to an AI or a person, the act of thinking first changes everything. It's the difference between using a tool and being used by one. The webinar will be available for those who missed it, but the real work—the institutional and personal work of learning to think alongside machines rather than be replaced by them—is just beginning.
Notable Quotes
Think before you ask. If you're throwing ideas out there before you've thought, the AI's thinking becomes your thinking and you don't get a choice.— Michael J. Jabbour, Microsoft AI Innovation Officer
The question is not whether we will embrace AI, but how we will do so.— Wendy Hensel, UH President
The Hearth Conversation Another angle on the story
Why does Jabbour keep coming back to this idea of asking the AI to teach rather than tell?
Because if you ask without thinking first, you've already surrendered the thinking part. The AI fills the void. You end up with an answer, but not understanding.
But isn't that how people use tools all the time? You don't have to understand how a calculator works to use it.
True, but a calculator answers a question you've already formed. With AI, the question itself is still being shaped. If you haven't thought it through, the AI's logic becomes the shape of your thinking.
So the real skill isn't learning to use AI—it's learning to think better?
Exactly. The technology is almost secondary. The skill is knowing what you actually want to ask, and why you're asking it.
Jabbour mentioned seventy percent of work could be automated. That's terrifying for students right now.
It would be, except he's not saying those jobs disappear. He's saying the nature of the work changes. You're not doing the routine part anymore—you're directing the outcome.
Which requires a different kind of person entirely.
A person who can think clearly about what matters, communicate it, and adjust when the AI shows you something unexpected. That's harder to teach than a skill, but it's also harder to automate.