In my recent research, I wanted to see if we could find a better way to predict these price spikes before they happen. I decided to set up a “race” between old-school math and modern Artificial Intelligence (AI).

The Contestants: Traditional Math vs. AI

I compared four different “tools” to see which one could guess Indonesia’s inflation most accurately:

• ARIMA & SARIMA: These are the “old-school” models. They look at past prices to guess future ones. SARIMA is a bit smarter because it knows prices often spike during certain seasons, like the Eid al-Fitr holidays.

• ARIMAX: This model is more advanced because it doesn’t just look at past prices. It also looks at outside factors like the BI Rate (bank interest rates) and Global Oil Prices.

• Multivariate LSTM: This is the AI “Deep Learning” model. It works a bit like a human brain, using a special “memory” to decide which economic information is important to keep and what to ignore.

What Drives the Prices?

To make the predictions as real as possible, I used data from 2015 to 2023, including:

• Interest Rates (BI Rate): Which changes how much people spend.

• Exchange Rates: How much the Rupiah is worth against the US Dollar.

• Oil and Commodity Prices: Because when global energy prices go up, our prices usually follow.

And the Winner Is...

After testing these models against real-world data, the AI (LSTM) won by a landslide.

The Scorecard (Lower is Better)

I used a score called MAPE (Mean Absolute Percentage Error) to see how far off the guesses were.

My AI model was much more precise because it can handle “non-linear” patterns—meaning it understands that the economy doesn’t always move in a straight, predictable line.

Seeing the Difference

In the graph above, the black line is the actual inflation. You can see my red line (the AI) stayed very close to it, even during the bumpy years of 2022 and 2023. The other models (the dotted lines) were often way off or too slow to react.

Why This Matters

This isn’t just about cool math; it’s about helping people. My research suggests that Bank Indonesia could use AI models like this as an “early-warning system”.

If we can predict price spikes more accurately, the government can make better decisions about interest rates at the right time. This helps keep the economy stable and protects everyone’s savings.

The future of managing our economy might just have a lot more to do with AI than we thought!

I pitted a Disciplined AI against an Emotional Human to see who would walk away with more money after a total market meltdown. Here is how it went down.

Phase 1: The Raw Reality

First, I needed a playground. I pulled a month’s worth of real Bitcoin data—745 hours of price movements—to use as my baseline.

What are you looking at? This is our “blank canvas.” It shows the real price of Bitcoin over 745 hours. It’s a bit of a roller coaster, but it gives us a realistic foundation to start our experiment.

Phase 2: The Players Enter the Room

In this digital world, I created two types of investors:

• The “Stoic” AI: This guy follows a strict recipe. If the price drops below its “fair value,” it buys. If it goes above, it sells. It doesn’t have a heart, and it definitely doesn’t watch the news.

• The “Panic” Human: This agent is just like us. It suffers from FOMO (buying when the price goes up because everyone is excited) and Panic (selling as soon as the price drops because it’s scared).

When you put them together, the market changes.

What are you looking at? Notice the blue line. That’s the “Hybrid Market” where both AI and humans are trading. Because the humans are constantly chasing trends, they actually make the price even more “jittery” and volatile than the original data (the gray line).

Phase 3: The 50% Nightmare

To really test them, I did something mean. At hour 400, I forced the market to crash by 50% instantly. This is the “Flash Crash” scenario.

As the price plummeted, the human agents did exactly what we all fear: they panicked. They sold everything at the very bottom just to “save” what was left. The AI, however, saw a massive discount. While the humans were running for the exits, the AI stayed calm and stuck to its plan.

What are you looking at? This is the “Moment of Truth.” See that dotted line? That’s the crash.

• The Red Line (Humans) sinks like a stone because they sold out in a panic.

• The Green Line (AI) stays much higher because it kept its cool and even bought more while things were cheap.

The Final Verdict: Why Boring Wins

When the simulation ended, the results were staggering. The humans lost a huge chunk of their net worth because they sold when things were at their worst. The AI protected its wealth and even came out ahead.

The lesson here is simple: The market doesn’t usually beat us; we beat ourselves.

The AI didn’t win because it was “smarter” or could predict the future. It won because it was boring. It had a plan, and it didn’t let a 50% drop change its mind. In a world of market chaos, being boring is actually the ultimate superpower.

So, the next time you see the market crashing, ask yourself: Are you acting like the green line or the red line?

You studied for years, mastered a profession, and expected stability in return. Society reinforced this belief: education equals security. The more technical your job, the safer you were supposed to be.

Then AI arrived—not loudly, not violently—but quietly and efficiently.

No mass unemployment overnight.
No dramatic collapse.

Just a slow, almost invisible shift:
tasks becoming smaller, judgment becoming optional, and value leaking away.

This project started with a simple observation:

Jobs today don’t disappear suddenly. They weaken first.

That weakness looks a lot like something we already understand very well in another field: evolutionary biology.

Thinking About Work Like an Ecosystem

In nature, species do not survive because they are popular or prestigious.
They survive because they fit their environment.

The modern labor market behaves the same way.

Technology changes the environment.
Jobs must adapt—or slowly lose relevance.

To study this, I treat each occupation as a species inside an economic ecosystem shaped by AI.

Two forces determine survival.

The Evolutionary Framework: Jobs as Species

In biology, survival depends on two forces:

1. Internal capability (fitness)

2. External pressure (environment)

I apply the same logic to labor markets.

1. Skill Complexity (s)

This represents how deeply human a job is:

• judgment

• creativity

• leadership

• problem framing

• contextual decision-making

High s means the job relies on abilities that are hard to automate.

2. Technological Pressure (t)

This captures how exposed a job is to:

• AI systems

• automation

• software replacement

High t means the job operates in environments where machines scale quickly.

Using thousands of occupations from O*NET, each job is positioned in a skill–technology space.

Measuring Labor “Fitness”

To translate evolution into economics, I define a Fitness Score (F).

In simple terms:

• Fitness is economic survival capability

• High fitness → stable demand and wages

• Low fitness → declining relevance

Jobs with strong human skills lose fitness slowly.
Jobs under heavy automation lose fitness rapidly.

The 2030 Labor Compass

The 2030 Labor Compass chart illustrates a neural network’s prediction of occupational survival, mapping jobs based on the tension between Skill Complexity and Technological Pressure. It highlights a clear divide: occupations in the "Safe Zone" leverage high human talent with low AI impact to maintain maximum fitness, whereas those in the "Dead Zone" face extreme extinction risk due to low skill requirements and high automation exposure. Ultimately, the data suggests that long-term career viability in 2030 will depend on shifting toward "Elite Tech" roles where AI augments rather than replaces complex human tasks, or remaining in highly specialized, low-exposure niches.

Modeling Change Over Time: Why Static Maps Are Not Enough

A snapshot shows today.
Evolution is about movement.

To model how jobs shift over time, I simulate a process I call AI Drift:

• skills gradually lose rarity

• technology pressure increases yearly

• jobs slide across the fitness landscape

This is modeled using an LSTM neural network, which is designed to learn sequences and trajectories.

LSTM Forecast: The Great Leftward Shift

The LSTM Forecast: The Great Leftward Shift 3D visualization provides a dynamic look at the evolutionary trajectory of the labor market, where faded diamond markers represent the predicted future positions of various occupations. By incorporating a third dimension—Economic Fitness—the model reveals a significant migration of jobs as they react to escalating technological pressures and shifting skill requirements. The transition from purple (dead) to yellow (safe) zones underscores a "fitness" landscape where roles that fail to adapt are squeezed out, while those that evolve toward higher complexity and strategic AI integration find stability. This "leftward shift" captures the inevitable movement of the workforce as it seeks equilibrium in an era defined by rapid automation and changing economic value.

What the Model Shows (Narrative Insight)

Most jobs drift:

• slightly downward (lower fitness)

• toward higher automation exposure

Only a small group resists this drift:

• jobs with exceptional complexity

• roles that integrate AI strategically

This is not a collapse.
It is a slow evolutionary filter.

Impact Summary: Winners and Losers by 2030

Here is a simplified summary of what the model predicts:

Category	                  Observation
Global trend	:     ~18% average decline in job fitness
High-risk occupations	:      30 roles enter danger zone
Surprising losers	:      Routine tech & IT roles
Stable winners	:      Engineering, energy, medical leadership
Key survival trait	:      Human complexity, not tools

One uncomfortable finding stands out:

Being “in tech” is no longer a guarantee of safety.

What This Means for Individuals

The old survival strategy:

• learn a technical skill

• master a tool

• specialize narrowly

The new strategy:

• build judgment

• combine disciplines

• understand systems, not just software

Tools expire.
Human complexity compounds.

Final Thought: Selection, Not Destruction

AI is not destroying work.

It is selecting for:

• adaptability over routine

• synthesis over execution

• evolution over stability

In the next decade, jobs will not disappear first.

They will fade.

And only those that evolve fast enough
will remain visible on the economic landscape.

Thank you for Reading :)

Seeing this made me think about my own younger sibling. I decided not to give her a phone yet. She is still in middle school, and I feel she is not ready to use AI. My main concern is that if he has AI, she might stop thinking carefully and depend too much on it. I learned about AI myself only when I started college, and by that time, I was more prepared to use it wisely.

Challenges in Rural Areas

In my village, few people really understand AI. Most parents and teachers do not know how powerful it can be or how it can affect learning. Some parents focus only on grades. They feel, “If the child can get good marks, it does not matter how.”

Because of this, some kids get used to shortcuts, and they do not practice problem-solving or critical thinking. This is a challenge in many rural areas: limited access to information about AI, few adults who can guide children properly, and different priorities between learning and results.

Technology Is a Tool, Not a Shortcut

This situation made me realize something important: technology is not bad, but it can be misused if people are not ready. Kids need time to develop basic thinking skills before AI can help them. Without these skills, AI can become a crutch rather than a tool.

For my sibling, I want her to learn how to solve problems, understand lessons deeply, and think independently. Only after she has a strong foundation should she start using AI to support learning. This way, AI will become a helpful tool, not a replacement for thinking.

How AI Is Used by Students Today

Recent global data shows that AI is now a common part of student life:

• Around 86–88% of students worldwide report using AI tools like ChatGPT for schoolwork (resourcera.com).

• In some studies, 57.6% of students use AI weekly, and 18.8% use it daily (mdpi.com).

• Students use AI for different school tasks — about 35% for generative tasks (like ChatGPT), 29% for writing support, and 23% for math help (gostudent.org).

The pie chart shows that most students use AI frequently, with 57.6% using it weekly and 18.8% using it daily. This means AI is no longer an occasional tool but a regular part of students’ learning routines. Such frequent use, especially at a young age, increases the risk of dependency, where students rely on AI for answers instead of developing their own problem-solving and critical thinking skills. In rural areas, where guidance about responsible AI use is limited, this pattern can weaken learning habits and reduce deep understanding. This data supports the idea that AI should be introduced carefully and gradually, after children have built strong foundations in thinking and learning independently.

The bar chart shows that students most frequently use AI for generative tasks (35%), followed by writing support (29%) and math help (23%), while other uses remain relatively low. This pattern suggests that AI is mainly used to produce answers or complete core academic tasks rather than to support deeper understanding or exploration. The high use of AI for generative and writing tasks indicates a shift toward convenience and speed, where students may prioritize finished outputs over the learning process itself. When this behavior starts early, especially without guidance, AI risks replacing essential skills such as reasoning, problem-solving, and independent thinking. This reinforces the argument that AI should be introduced carefully and used as a learning aid only after students have built strong foundational skills.

Looking at the Bigger Picture

Many children in rural areas face the same challenge. Introducing AI too early can sometimes harm learning habits instead of helping. Parents and teachers need to understand not only how AI works but also when it is right to let children use it.

Reflecting on this, I feel it is our responsibility to guide the younger generation carefully. AI is here to stay, and it can change education and daily life dramatically. But its value depends on how and when we use it. Learning to think first, then use AI wisely, is the most important lesson I can give to my sibling.

Conclusion: Think First, Use Later

I hope one day, more families in villages like mine will understand this. AI can be a powerful helper if children are ready, but if they are not, it can become a shortcut that stops them from learning important skills. Technology alone cannot replace curiosity, patience, and careful thinking.

I end this note with a simple thought: teach children to think first, and introduce AI later. This is my small way of preparing the next generation for a future where technology is everywhere.

“Perempuan ada di awal, tapi jarang sampai ke puncak.”

Ketika kita bicara tentang kesetaraan gender di era teknologi canggih seperti AI, banyak yang mengira masalahnya ada di akses awal ke pendidikan. Padahal, data dari World Economic Forum dan LinkedIn menunjukkan bahwa tantangan sebenarnya muncul justru setelah perempuan lulus dan mulai bekerja.

Fenomena "Drop-off": Dari Kampus ke Kantor

Menurut laporan Gender Parity in the Intelligent Age (2025), ada tren yang konsisten:

• Jumlah perempuan lulusan STEM (sains, teknologi, teknik, dan matematika) terus meningkat.

• Namun, ketika masuk ke dunia kerja, angka partisipasi mereka langsung anjlok.

Sebagai contoh: lulusan perempuan tahun 2017 mencakup 35.5% dari total lulusan STEM, tetapi hanya 29.6% dari mereka yang masuk ke pekerjaan STEM di tahun berikutnya.

Penurunan partisipasi perempuan setelah lulus dari pendidikan STEM

Sumber: World Economic Forum & LinkedIn Economic Graph Research Institute, 2025.

Faktor-Faktor Penyebab

1. Bias dan stereotip industri

   Perempuan masih sering dianggap “kurang cocok” untuk bidang teknis atau posisi pemimpin di bidang teknologi.

2. Kurangnya dukungan karier awal

   Mentoring, jaringan profesional, dan program pengembangan karier jarang dirancang dengan mempertimbangkan kebutuhan perempuan.

3. Minimnya role model

   Perempuan yang melihat kurangnya figur sukses di bidang STEM cenderung merasa “tidak ada tempat” untuk mereka.

4. Lingkungan kerja yang tidak inklusif

   Mulai dari mikro-agresi hingga kurangnya kebijakan kerja fleksibel, banyak tempat kerja belum cukup ramah gender.

“Drop to the Top”: Semakin Tinggi, Semakin Sedikit

Masalahnya tidak berhenti di level entry. Ketimpangan semakin besar di level manajerial dan eksekutif:

• Di tahun 2024, hanya 12.2% dari pemimpin C-suite di bidang STEM adalah perempuan, meskipun mereka mengisi 24.4% posisi manajer.

Ini menciptakan siklus yang sulit diputus: tanpa pemimpin perempuan, semakin sedikit perempuan muda yang melihat jalan karier yang mungkin ditempuh.

Representasi perempuan dalam jenjang karier STEM

Sumber: World Economic Forum & LinkedIn Economic Graph Research Institute, 2025.

AI adalah Peluang – Jika Kita Bertindak Sekarang

Berita baiknya: bidang AI (Artificial Intelligence) masih tergolong baru dan belum “terlalu kaku” oleh ketimpangan gender. Dengan strategi yang tepat, kita bisa:

• Meningkatkan skilling & reskilling perempuan di bidang AI.

• Membuat pipeline karier yang adil sejak pendidikan hingga kepemimpinan.

• Menyusun kebijakan perusahaan dan publik yang menghapus hambatan struktural.

Jika kita ingin dunia AI yang adil, kita harus mulai dengan mempertahankan talenta perempuan yang sudah ada. Investasi pada dukungan karier, lingkungan kerja inklusif, dan kebijakan progresif bukan hanya isu keadilan — ini soal masa depan inovasi dan ekonomi global.

Seluruh data dan analisis dalam artikel ini bersumber dari laporan "Gender Parity in the Intelligent Age", diterbitkan oleh World Economic Forum pada Maret 2025, bekerja sama dengan LinkedIn Economic Graph Research Institute. Laporan ini membahas secara komprehensif bagaimana kesenjangan gender di bidang teknologi dan kecerdasan buatan (AI) terbentuk, serta strategi untuk mewujudkan transformasi digital yang lebih inklusif.

Tren Global yang Mempengaruhi Indonesia

Lima tren makro yang paling berpengaruh terhadap transformasi pasar kerja global juga relevan bagi Indonesia:

• Perubahan teknologi (60% perusahaan global memperkirakan ini akan mentransformasi bisnis mereka).

• Kenaikan biaya hidup (50%).

• Mitigasi perubahan iklim (47%).

• Adaptasi terhadap perubahan iklim (41%).

• Perubahan demografi seperti populasi usia kerja yang tumbuh di negara berkembang, termasuk Indonesia (25%).

Secara khusus, perluasan akses digital dan adopsi teknologi seperti AI, big data, dan robotika menjadi pendorong utama perubahan pekerjaan di banyak sektor

Macrotrends driving business transformation

Source : World Economic Forum

Pekerjaan yang Tumbuh dan Menurun

Berdasarkan prediksi global hingga tahun 2030, diperkirakan:

• 170 juta pekerjaan baru akan tercipta (14% dari total pekerjaan saat ini).

• 92 juta pekerjaan akan hilang (8% dari total pekerjaan), menghasilkan pertumbuhan bersih 78 juta pekerjaan.

Pekerjaan yang tumbuh cepat:

• Big Data Specialist

• FinTech Engineers

• AI & Machine Learning Specialist

• Software and Applications Developers

• Environmental Engineers

• Renewable Energy Engineers

• Delivery Drivers

• Nursing Professionals

• Tertiary Education Teachers

  

Fastest-growing jobs, 2025–2030

Source : World Economic Forum

Pekerjaan yang menyusut:

• Data Entry Clerks

• Administrative Assistants

• Bank Tellers

• Cashiers dan Ticket Clerks

• Legal Secretaries

• Printing Workers

  

Fastest-declining jobs, 2025–2030

Source : World Economic Forum

Sektor seperti pertanian, pengiriman barang, konstruksi, dan pendidikan akan mengalami pertumbuhan besar secara absolut—hal ini relevan dengan struktur ekonomi dan tenaga kerja Indonesia

Kesenjangan Keterampilan: Ancaman dan Peluang

Sekitar 63% perusahaan mengidentifikasi skill gaps sebagai hambatan terbesar dalam transformasi bisnis. Sebanyak 59 dari 100 pekerja membutuhkan pelatihan ulang atau peningkatan keterampilan sebelum 2030.

Keterampilan yang paling dibutuhkan:

• Analytical thinking

• Ketahanan, fleksibilitas dan kelincahan

• Kepemimpinan dan pengaruh sosial

• Literasi teknologi, AI dan big data

• Cybersecurity dan jaringan

Sebaliknya, keterampilan manual seperti ketahanan fisik dan ketelitian menurun dalam permintaan.

Top growing and declining skills, 2025–2030

Source : World Economic Forum

Peluang di Sektor Hijau dan Digital

Indonesia memiliki potensi besar dalam dua sektor kunci:

• Pekerjaan hijau: Renewable Energy Engineers, Environmental Engineers, dan tenaga kerja kendaraan listrik—didorong oleh komitmen iklim dan pertumbuhan industri energi terbarukan.

• Ekonomi digital: Developer, Analis Data, dan Spesialis UI/UX menjadi pekerjaan paling cepat tumbuh karena peningkatan akses digital dan adopsi teknologi AI.

Namun, laporan ini juga mencatat bahwa permintaan keterampilan hijau meningkat 22%, sementara pertumbuhan pekerja dengan green skills hanya naik 12%—menunjukkan adanya kesenjangan signifikan yang perlu dijembatani melalui kebijakan pelatihan nasional

Kesimpulan

Laporan Future of Jobs Report 2025 menunjukkan bahwa pasar kerja Indonesia akan berada di tengah arus besar transformasi global. Dengan populasi usia kerja yang terus bertumbuh, Indonesia memiliki peluang besar untuk memimpin dalam penciptaan pekerjaan baru—terutama di sektor hijau dan digital. Namun, peluang ini hanya akan maksimal jika dibarengi dengan peningkatan keterampilan, kesiapan teknologi, dan dukungan kebijakan yang inklusif.

Penting bagi seluruh pemangku kepentingan—pemerintah, dunia usaha, lembaga pendidikan, dan masyarakat—untuk bersama-sama menyiapkan tenaga kerja Indonesia menuju tahun 2030 dengan kesiapan keterampilan yang relevan dan berdaya saing global.

Data dan analisis dalam artikel ini bersumber dari:
World Economic Forum, Future of Jobs Report 2025.
Diakses dari: www.weforum.org