Why Do We Need Intelligent Systems?

Bowman: Open the pod bay doors, HAL. 
HAL 9000: I am sorry, Dave. I am afraid I can not do that. 
-- 2001: A Space Odyssey

System Data Wisdom Process Optimization

One Movie Changed Everything

A movie from my childhood left an indelible mark on my mind. I am still carrying its resonance with me today. The protagonist of that film was ‘HAL’, a sentient computer. You guessed it right — the film was 2001: A Space Odyssey. In that movie, we witnessed the intricate tension between humanity and machines that can ‘think’.

Why Do People Want Intelligent Machines?

For a century, our fascination with intelligent machines has manifested in many ways. Our yearning and excitement for personalities like Iron Man’s JARVIS, the Terminator, or Knight Rider’s KITT represent a profoundly powerful idea. Whether machines can truly think like humans right now, or whether such thinking machines are being built today or not — what matters most is that humanity refuses to let go of this vision.

Machines Excel at Specialized Tasks

To solve the monumental challenges ahead — curing diseases, DNA sequencing, mitigating natural disasters, exploring distant galaxies — we need intelligent systems standing by our side. Humans need them for their own survival and progress. Consider this: the vast majority of global business transactions now run through credit cards. During last year’s AliExpress 11:11 festival, $1 billion in sales took place in just the first 68 seconds. The rest is history. This means that processing such colossal volumes of credit card data for real-time fraud detection is simply impossible for thousands of human beings alone. We must enlist the support of intelligent systems.

Our Dependence on Machines

I am certainly no exception. My own heavy reliance on machines for everyday household chores is a clear reflection of this. I operate on the premise that our purpose — human purpose — lies elsewhere. We must delegate routine tasks to machines so they can absorb our operational burden. Naturally, we want machines capable of thinking as we do. For routine conditions, an ‘if-this-then-that’ approach (somewhat like IFTTT) works well. But when confronted with the unknown, machines must make reasoned decisions just like humans. We already observe this in medical diagnostics, autonomous vehicles, long-haul flights, and on-demand cloud resource allocation in data centers, where rule-based systems maintain smooth operations most of the time. In essence, we want machines to stand beside us as collaborative partners.

Massive Population, Process Optimization

Consider the time it currently takes to receive a government service. If ‘thinking’ machines join hands with human workers, the processing turnaround for permits, trade licenses, driving licenses, passports, and utility connections will dramatically drop without question. We marvel at how credit card verification happens online in seconds; governmental workflows are often far simpler and largely process-driven. To eliminate human bias — deciding whom to serve and whom to deny — we urgently need intelligent systems of this caliber.

Scaling Public and Private Systems

Our work must happen at scale — on an immense level. To deliver services to 160 million people, processes must be automated. While conventional legacy software functions today, it will inevitably fall behind in the near future. My central point is straightforward: alongside traditional methods, intelligent systems must step in to eliminate societal bottlenecks and inconsistencies. Intelligent systems will streamline public service delivery, dramatically cut processing times, assist doctors with precise diagnostics, and ultimately serve humanity.

National Policy Dashboards

Never doubt that a small group of thoughtful, committed citizens can change the world; indeed, it’s the only thing that ever has.

— Margaret Mead

Can you tell how a truly efficient company operates? You guessed it right: through profit-and-loss spreadsheets. They account for every single taka of investment. Where profits are generated and where losses occur can be tracked directly from those spreadsheets. For a small company, keeping track is straightforward. As an organization expands, the challenge compounds: hundreds of divisions, each maintaining its own ledgers. Distinguishing which unit yielded a 2% profit while another suffered an 80% loss within the same row becomes arduous. That is precisely why Chief Executives of multi-million and multi-billion dollar enterprises rely on executive ‘dashboards’ — providing high-level situational awareness for making informed decisions. A broad perspective appears right there on the screen, while granular details remain accessible with a single click.

How does a successful enterprise function? Consider Grameenphone. The Chief Executive Officer has a high-level executive dashboard. Company profitability, departmental KPI progression, distance from forward targets — everything converges onto this single dashboard. Every analytical tool needed to make an informed decision is placed at their fingertips. Alongside this, the dashboard reflects the strategic benchmarks of Telenor — the governing philosophy of the parent company. Every enterprise operates on an underlying philosophy; to capture that philosophy, essential policies cascade throughout the organization in a structured flow. Furthermore, every component inside is interconnected. A single SIM card sale in Bangladesh can register its impact on the global dashboard. The data points may seem tiny, yet everything is linked by a unified thread. When government policy shifts — such as a rise or drop in SIM taxation — the outcome immediately reflects there. Mid-level managers also operate with dashboards tailored to their scope, enabling informed decision-making at every operational tier.

What would happen if a government operated with such a dashboard? Data holds enormous power; it can penetrate deep beneath the surface. Every event in the world is connected to another — working with data makes this vividly clear. From a teenage girl’s suicide driven by extreme poverty to fatalities from road traffic accidents, every ripple can surface on that national dashboard. Where the distribution of Vulnerable Group Feeding (VGF) cards encountered field-level blockages would also appear clearly. We might not even require exhaustive micro-details right now; village-level data points are already sufficient. The head of government could uncover crucial underlying realities — before crises escalate. Singapore’s “Smart Nation” and the United Arab Emirates’ “Smart Gov” dashboards have already reached unprecedented heights, where nearly their entire public service delivery platform operates seamlessly online. We too are moving forward step by step. That transformation can begin with this very dashboard.

Can a government be run this way, like an enterprise? Absolutely. Many corporations today manage their Corporate Social Responsibility (CSR) initiatives with a civic responsibility akin to governments — driven by accountability to people, where immediate profit is not the sole objective. As noted earlier, nations from Singapore outward are experiencing firsthand the transformative power of data. Frankly speaking, after attending a regulatory training program in Singapore, my entire perspective expanded. In those nations, integrating data into policy formulation saves enormous sums of public money — curbing waste and preventing costly missteps. Data empowers leaders to make timely decisions, and a single timely intervention can save fortunes; otherwise, project expenditures escalate uncontrollably day after day.

What would such a dashboard look like? A data-driven national policy dashboard for the Prime Minister, Members of Parliament, and senior civil servants — customized for each domain. Over 200 distinct data streams, encompassing ministries and field agencies, would feed continuous real-time data into the system. Correlating these data streams through machine learning would reveal genuine public needs, spotlight core underlying issues, and track the progress or implementation gaps in each MP’s constituency. Where initiatives are stalled, what operational roadblocks exist, and what internal irregularities persist — everything would surface on our dashboard. Consequently, decisions would become genuinely informed — grounded in evidence rather than conjecture. Our observations show that data-driven governance dashboards have already been implemented in Indian states such as Andhra Pradesh, Odisha, Jharkhand, and Maharashtra with the collaboration of private technical organizations.

I have served in government for nearly 30 years. The utilization of data is expanding within government circles, applied in select strategic areas, and a National Open Data Initiative has been inaugurated. Yet, that alone is not enough. Living and working with data daily, I intimately understand its vital importance. A strategic policy dashboard for high-level decision-making continuously turns in my mind. My PhD proposal centers on this exact mission: “National Decision Support System for Bangladeshi Policymakers from Anonymous Large Scale Mobile Phone Data”. Machine learning forms a substantial component here. The framework operates across three pillars: (1) building a National Decision Support System, (2) tailored for national policymakers, and (3) driven by large-scale data points. Because I am working on this independently, I scoped it into a concrete framework subject to the availability of anonymized mobile phone data. With a broader array of data inputs, it will function at an even larger scale. Everything is achievable — through artificial intelligence.

Starting with Small-Scale Intelligent Machines

Much of how we humans reason has already been imparted to machines. Whether an elevator continues accepting calls up to a given time, whether it stops at our floor if calls are queued above, whether it bypasses external calls when overcrowded, or in a multi-elevator bank, which car will arrive at your floor in the shortest time based on proximity and load calculations — these are all routine everyday computations.

Software Is Eating the World

People often ask me: why should we learn machine learning or deep learning? My answer is simple: this is the defining field transforming the entire planet. This emerging discipline is continuously exploring what remains unattainable through conventional software. Look around you right now — the most prominent enterprises are software companies: bKash, Pathao, Daraz, and many others. Yet, machine learning and deep learning will fundamentally disrupt and assimilate traditional software companies.

“AI Will Eat Software”

Do a quick search: “Software is eating the world, AI will eat software”. Whatever challenges stand before you, exploring problem spaces through machine learning allows you to solve virtually any category of problem — as I observe all around me. This exploration is deeply intellectual: I continually encounter novel problems whose only viable solution lies in machine learning.

We Must Elevate Our Skillsets

Because thinking machines possess synthetic intelligence, our central dialogue revolves around artificial intelligence. However, the foundational concepts of teaching machines are still clouded in ambiguity for many. That is precisely the reason behind this series of books, videos, and discussions.

I believe in a timeless truth: “If you build it, they will come.” Once we cultivate skilled human capital, the economic transformation will inevitably take root here. We need a vast pool of skilled individuals to build intelligent systems. Software must now be endowed with intelligence, and that requires willpower, dedication, and intellect.


Updated