L

Hi, I'm Laura — curious by nature, researcher by choice.

DBAI candidate at PolyU Hong Kong. I write about AI, research, responsible tech, and the messy, fascinating process of figuring things out.

DBAI @ PolyU HK Transformative AI Responsible AI Always learning
About Me.
Random thoughts on the journey
All posts →

I'm someone who ended up in a DBAI programme partly out of curiosity and partly because I couldn't stop asking questions that didn't have tidy answers. I work with AI every day — and the more I work with it, the more I realise how little we collectively understand about where it's taking us.

This corner of the internet is where I think out loud. No polished takes, no pretending I have it figured out. Just honest reflections from someone in the middle of it.

When I'm not reading research papers or arguing with AI systems, you'll find me somewhere between Hong Kong and Hanoi, perpetually caffeinated.


Journal
Will governance help AI adoption, or quietly stop it?
I want to understand whether AI governance gives people confidence to move, or gives them one more reason to do nothing.
June 2026 Read →
Journal
The most dangerous word in innovation is "interesting"
The dangerous projects are the ones everyone finds "interesting" — everyone nods, everyone smiles, and then absolutely nothing happens.
June 2026 Read →
Journal
Curiosity is a very expensive hobby
It starts with one article and ends at midnight with three new questions, less certainty, and no improvement to my sleep schedule.
June 2026 Read →
Journal
AI has made me significantly less impressive
AI has not removed the value of work. It has shifted the value from execution toward judgment, deciding what matters, and knowing what to do next.
June 2026 Read →
Research Journey.
Notes from my DBAI at PolyU Hong Kong
All posts →
Research
Human oversight needs more than an approval button
Decide which parts of a workflow AI should handle before deciding how much authority to give it.
October 26, 2026Read →
Research
An explanation is not the same as oversight
A convincing explanation can still hide how a model reaches its answer. Effective review requires something people can actually inspect.
October 25, 2026Read →
Research
A little control can make AI easier to use
People were more willing to use an imperfect algorithm when they could adjust its forecasts — even slightly.
October 24, 2026Read →
Research
Hidden AI reasoning is not automatically private
A new paper shows that encrypted reasoning traces can become a privacy and security problem when they move across sessions, users, and models.
August 17, 2026 Read →
Reflections.
On responsible AI — the questions that keep me up at night
All posts →
Responsible AI
People judge AI like they judge people
AI adoption is not only about capability. People also judge whether the system feels warm, fair, honest, and safe to rely on.
July 26, 2026 Read →
Responsible AI
AI leadership still needs someone human in the room
AI can help leaders prepare, analyse, and evaluate, but leadership is not only information processing with better formatting.
July 26, 2026 Read →
Responsible AI
AI is not always the right front desk
AI can make service faster, but the research shows people still care deeply about when a human is present, especially when the moment feels personal.
July 26, 2026 Read →
Responsible AI
Human-AI teams need instructions, not optimism
Human plus AI sounds powerful, but the research keeps reminding me that collaboration does not become useful just because both parties are present.
July 26, 2026 Read →
Responsible AI
AI trust travels through people
We keep trying to solve AI trust with better models, but trust often travels through people first.
June 27, 2026 Read →
Responsible AI
Human-AI collaboration needs design, not slogans
Putting AI into a workflow is easy. Designing the relationship between human judgment and machine output is the hard part.
June 27, 2026 Read →