MONASH UNIVERSITY = Women in AI: Breaking Barriers and Building Futures panel event [ Part 2 ]

Baik itu pariwisata, smart meter, smart city—semuanya adalah sistem. Kami membangun sistem komunikasi, rekam medis elektronik, hingga AI masa kini. Semuanya berada di titik persimpangan tentang bagaimana kita menerapkan sistem yang usable untuk masyarakat dan organisasi. Jadi meskipun pekerjaan saya sekarang disebut AI, pada dasarnya tetap tentang bagaimana manusia dan organisasi mengadopsi teknologi baru dan membuatnya bermanfaat. Itulah perjalanan saya hingga masuk ke dunia AI.

Kalau kita bicara masa lalu, sebenarnya itu sudah lewat—yang terpenting adalah masa depan. Tapi, orang bisa mendapatkan wawasan dari masa lalu saya. Ketika kecil, saya tidak punya mimpi menjadi insinyur. Saya hanya tahu bahwa saya suka matematika. Matematika waktu SD dulu sangat mudah—tidak seperti sekarang yang sulit sekali.

Di keluarga saya, tidak ada yang sekolah di ITB atau menjadi engineer. Mayoritas adalah guru. Saya adalah cucu pertama yang masuk teknik di ITB. Mengapa saya masuk teknik? Sebenarnya karena ayah saya. Beliau seorang akuntan, dan mungkin melihat bahwa pekerjaan akuntansi akan banyak otomatisasi. Jadi beliau berkata, “Kalau kamu masuk Informatika ITB, ayah akan dukung.”

Saat itu saya tidak tahu apa itu informatika. Ketika masuk, saya kaget: “Ini apa sebenarnya?” Seperti yang tadi Rena bilang, seolah-olah Tuhan yang menuntun kita ke sana. Saya tidak punya visi khusus. Saya hanya menjalani, belajar, mengikuti arah hidup.

Tahun keempat semua mahasiswa harus membuat tugas akhir. Saya bingung mau apa. Saya tidak suka keamanan siber, tidak suka jaringan. Tapi saya tertarik penerjemahan mesin. Jadi saya membuat sistem penerjemahan sederhana—tanpa machine learning, karena itu 27 tahun lalu. Saya membuat penerjemah Indonesia–Inggris berbasis aturan. Dari sana semua penelitian dan mata kuliah yang saya ajarkan selalu berhubungan dengan NLP dan AI. Seolah memang itu jalan hidup saya.

Sekarang saya sudah lebih dari 40 tahun, dan fokus saya adalah masa depan—apa yang bisa saya lakukan dengan sisa hidup ini. Saya ingin mengajar AI kepada lebih banyak orang dan membuat produk yang bermanfaat.

🔎 Tentang Proyek AI yang Sukses

1. Perspektif Akademisi

Menurut pengalaman saya, proyek AI yang sukses adalah proyek yang tetap digunakan sampai sekarang. Kalau tidak dipelihara, modelnya akan rusak dan tidak terpakai. Di berbagai proyek (baik di ITB maupun di perusahaan rintisan seperti Prosa.ai), saya melihat bahwa proyek sukses adalah ketika:

  • Pengguna atau institusi bisa memelihara sistem itu sendiri.

  • Prosesnya mengikuti etika yang benar—pengumpulan data, privasi, dll.

2. Perspektif Industri

Di industri, ada dua tipe proyek:

(a) Proyek berbasis produk
Kesuksesan berarti produknya bekerja sesuai janji, unggul di benchmark, dan bisa diandalkan.

(b) Proyek riset dalam organisasi besar
Kesuksesan bisa berarti:

  • Menemukan area produk yang bisa ditingkatkan.

  • Menemukan area di mana AI tidak cocok digunakan.

  • Memberikan jalur pertanyaan dan riset yang tepat.

Kadang, kesuksesan adalah menyimpulkan bahwa AI tidak perlu diterapkan pada kasus tertentu—misalnya masalah sebenarnya adalah infrastruktur atau ekonomi.

3. Perspektif Pemerintah

Tugas utama pemerintah adalah memberikan layanan publik.

Jadi, proyek AI yang sukses adalah:

  • Yang mempermudah, bukan mempersulit.

  • Yang benar-benar dibutuhkan dalam alur layanan.

  • Yang tidak ditambahkan hanya karena tren atau tekanan “harus pakai AI.”

Kadang pemimpin hanya ingin terlihat modern, sehingga memaksa penggunaan AI. Akhirnya menambah masalah, bukan menyelesaikannya. Prinsipnya:

“Kalau AI tidak membuat layanan lebih baik, jangan dipaksakan.”

4. Perspektif Kebijakan & Ekonomi

Kesuksesan juga bisa berarti:

  • Membangun fondasi data yang baik untuk masa depan—even jika manfaatnya baru terlihat 3–5 tahun kemudian.

  • Disiplin dalam proses: setiap proposal AI harus menjawab:

    • Siapa yang akan diuntungkan?

    • Seberapa besar manfaatnya?

    • Apa dampak ekonominya?

Whether we're talking about tourism attractions, smart meters, or smart cities — they are all systems. We build communication systems, electronic medical records, and now today’s AI systems. All of them sit at the intersection of one central question:

How do we design and deploy technology that is usable in real organizations and real communities?

So even though I call the work I do today “AI,” at its core it is still about understanding:
How do people and organizations adopt new technologies, and how do we make those technologies truly useful?
That’s essentially my journey into AI.

Looking Back – How It Started

If we talk about the past, well… it’s already in the past. The future is what matters. But I think sharing a bit of the past can offer insights.

When I was young, I had no dreams of becoming an engineer — none at all.
I only knew that I loved math. And the math back in my primary school days was very simple, nothing like the very difficult math kids learn today. It was just addition, multiplication — basic things. So that was the only subject I liked.

No one in my family studied engineering. Many of them became teachers. I was the first grandchild to ever study engineering, and honestly, it happened because of my father. He was an accountant, and I think he started seeing automation appearing in the accounting world. Maybe he sensed that the job might change or even be replaced someday.

So he told me,
“If you go to Informatics at ITB, I will fully support you.”
At that time, I didn’t even know what “informatics” meant. I had absolutely no idea. So when I entered Informatics, I was shocked — “What is this? What have I gotten myself into?”

And just like Rina said earlier, I feel like AI found us, not the other way around. I didn’t dream of it. I didn’t plan it. People are usually driven by vision or mission — but I simply followed the path that opened in front of me.

Another thing that shaped me was seeing how many of my relatives did not study seriously when they were young. Because of that, they couldn’t choose good majors or opportunities later in life. That stayed with me. So I pushed myself when I got to ITB.

My First Encounter With AI

In my fourth year, everyone had to choose a final project. I didn’t like cybersecurity, I didn’t like networking — so I didn’t know what to do.

But I have always liked machine translation.
So I decided to build one.

It was purely rule-based — no machine learning at all — because this was 27 years ago. It was a long time ago. I built my own machine translation system for Indonesian–English.

After that, everything in my academic journey aligned with that direction:
my research, my courses, my teaching — all centered around natural language processing and artificial intelligence.
It became clear that this was simply the path meant for me. God placed me here.

Now I’m over 40, and at this stage of life, the most important thing is the future. I already know a little about AI — not everything, but enough — and for the rest of my life, I want to do something meaningful:
to teach AI to many people, and to create AI products that will truly help society.
That is where my focus is now.

Moderator Transition

Thank you to all our speakers — it’s fascinating to hear that you don’t need an IT or computer science background to end up in AI. Your stories really highlight that.

Now I want us to zoom out a bit.
Let’s talk about the broader landscape.

✨ Question: What does a successful AI project look like in your sector?

(Academia, Industry, Government)

Speaker 1 (Academia & Startup Experience)

In my experience, a successful AI project is one that continues to be used over time.

If we don’t maintain a model, eventually it becomes useless.
During my years at ITB and later at Prosa.ai, I saw many AI projects delivered to companies or government institutions. The successful ones were the ones they could maintain themselves — meaning:

  • they had good knowledge transfer

  • they had proper data practices

  • they followed ethical standards in data collection and usage

When those elements exist, the AI system can continue to provide value long after it is deployed. That’s what success looks like to me.

Speaker 2 (Industry / Big Tech Research)

In industry, I see two types of AI projects:

1. Product-led AI projects

These are easy to evaluate.
A successful project = a successful product.
It performs well in benchmarks, delivers on its promises, and provides clear value.

2. Research-driven AI projects

This is what we do in the research arm of a large organization.

Here, “success” may look different.
Sometimes the project doesn’t lead to a new product — sometimes it identifies where current products are failing. That insight is still extremely valuable.

If we can map out:

  • what’s not working,

  • why it’s not working, and

  • how it can be improved,

then that is a successful AI project.

3. Societal & organizational readiness assessments

Another type of project looks at society, users, and organizations.
Sometimes the most responsible conclusion is:
“AI is not the answer here.”

Maybe the issue is:

  • infrastructure

  • economic conditions

  • or poor data foundations

If deploying AI won’t solve the problem, we step back.
Knowing when not to deploy AI is also success.

Speaker 3 (Government Perspective)

In government, our primary duty is to provide services:

  • services for citizens

  • services for businesses

  • services for other government agencies

  • services for our own employees

So when we introduce AI into public services, the key question is:
Does AI actually make things easier for the user?

Many leaders feel pressured — they don’t want to be left behind, so they push AI into everything. But sometimes this ends up creating new problems instead of solving old ones.

For example, we create a platform, but then we add layers of unnecessary bureaucracy inside that platform. So instead of simplifying, we complicate.

Therefore, before implementing AI, we must return to our mission:
If AI truly helps the service, we use it. If not, we don’t force it.
Not every government task needs AI.

Speaker 4 (Economic & Strategic Perspective)

It might be too early to talk strictly about ROI, but I always remind people:
everyone needs to share a piece of the pie — we must support our families and contribute to the country’s economy.

So skepticism about AI is healthy.
But like I said earlier, I didn’t choose AI — the opportunities opened, and I walked through the doors. And once we walk through, we need to capitalize on those opportunities responsibly.

Sometimes success doesn’t look measurable.
For example, maybe we can’t do full AI today — but we can teach someone how to collect data properly. Someone else may reap the benefits 3–5 years later. That’s still success.

But I also believe in discipline.
Anyone proposing an AI project should answer:

  • Who benefits?

  • How much value does it create?

  • What problem does it solve?

  • Is this the right moment to introduce AI?

To me, that is how we build meaningful AI impact.


Comments

Post a Comment

Popular Posts