Introduction: A Question of Limits
Optimization can serve humanity enormously, but only up to the point where it stops respecting two boundaries: the quality of what is produced, and the dignity and safety of the people who produce and use it. Beyond those boundaries, the pursuit of profit no longer creates wealth. It transfers wealth, and it transfers risk, from the many to the few.
We are living through a technological mutation that moves, figuratively, at the speed of light. Robotics, artificial intelligence and automated decision systems now reach into factories, hospitals, banks, schools and homes. Tasks that once required years of human training can be performed in seconds by a machine that never sleeps, never tires and never asks for a raise.
This power is neither good nor evil in itself. A hammer builds a house or breaks a window depending on the hand that holds it. The same is true of an algorithm. The real question is not whether we should optimize, but how far, for whom, and under whose control.
In this essay I argue a simple thesis. As long as technological change remains under human control, guided by law, ethics and education, humanity can reap immense benefits from it. If it escapes that control and falls into the hands of a small group driven only by greed, it can become harmful, unproductive, and even a path toward a new form of servitude and self-destruction.
What Optimization Really Means
As a mathematics teacher, I learned that no serious optimization problem has an objective function without constraints. A student asked to maximize the area of a field with a fixed length of fence must respect the fence. Remove the constraint and the problem becomes meaningless: the answer is simply “infinity”.
Business is no different. A company may legitimately seek to maximize profit, but it must do so subject to constraints that cannot be negotiated away. We can write the problem in the language every calculus student knows:
Maximize P(x), subject to: Q(x) ≥ Q min, H(x) ≥ H min, S(x) ≥ S min
Here P is profit, Q is the quality of the product, H is the well-being of the human beings involved, workers and customers alike, and S is the safety and stability of society. The minimum values are not suggestions. They are the walls of the problem.
The danger of our time is that powerful tools make it tempting to quietly erase those walls. When a machine can cut costs by thirty percent overnight, the pressure to ignore quality, safety and human consequences becomes enormous. The mathematics then turns into a caricature: maximize profit, full stop.
There is also a subtler trap, known among economists as Goodhart’s law: when a measure becomes a target, it ceases to be a good measure. A hospital optimized only for the number of patients seen per hour may treat each one worse. A call centre optimized only for call length may solve fewer problems. An AI system optimized only for clicks may spread anger and falsehood, because anger and falsehood attract clicks.
True optimization, therefore, is not the maximization of one number. It is the search for the best balance among several values that matter. Efficiency is a virtue. The obsession with a single figure, at the expense of everything else, is not.
The First Limit: The Quality of the Product
Quality is the first wall that profit-seeking must not break. A product is not merely a source of revenue; it is a promise made to the person who buys it. Every cost cut that silently weakens that promise is a form of borrowing against the future.
Automation, used wisely, actually raises quality. Robots weld with a precision no human hand can match. Machine vision detects defects invisible to the eye. AI systems can help a doctor notice a shadow on an X-ray that might otherwise be missed. In these cases, optimization and quality move together, and everyone gains.
The problem begins when automation is used not to improve the product but to remove the people who guarantee it. Inspectors are dismissed because “the software checks everything”. Engineers are pressed to ship faster because the competition is moving. Customer service is replaced by a chatbot that cannot understand a real problem.
The aviation industry offers a painful lesson. The two Boeing 737 MAX crashes in 2018 and 2019, which killed 346 people, were linked to an automated flight-control system, to pressure on schedules and costs, and to weaknesses in oversight and pilot training. The technology was not the only culprit. The deeper failure was a culture in which speed and cost were allowed to outrank safety.
The Japanese automaker Toyota has long expressed the opposite philosophy with the word jidoka, often translated as “automation with a human touch”. Machines stop automatically when something goes wrong, and any worker on the line may halt production to fix a defect. The machine serves the quality of the work; it does not replace human judgment about it.
The lesson is clear. Optimization should reduce waste, not reduce care. When a company saves money by lowering quality, it is not optimizing; it is spending its reputation, and sometimes human lives, to inflate a short-term figure.
The Second Limit: The Human Factor
The second and most important wall is the human being. An economy exists to serve people, not the reverse. Any optimization that treats human beings merely as costs to be eliminated has forgotten its own purpose.
The human factor has at least three dimensions. The first is work. For most people, a job is not only an income; it is a source of identity, structure and dignity. When automation destroys jobs faster than society can create new ones, the result is not efficiency but social fracture, anxiety and lost talent.
The second dimension is judgment. Machines are excellent at calculating; they are far weaker at understanding context, values and exceptions. A loan algorithm may reject a perfectly honest applicant because of a pattern in the data. A sentencing tool may repeat the prejudices hidden in past decisions. Where lives, freedom or livelihoods are at stake, a human being must remain responsible for the final decision.
The third dimension is skill. When people stop practising a craft because a machine does it for them, the knowledge slowly disappears. Pilots who rely too heavily on autopilot can lose their manual flying skills. Students who let software solve every equation never learn to reason. In my own classroom, I used interactive technology to teach senior mathematics, but always as a means to deeper understanding, never as a substitute for thinking.
None of this means we should refuse automation. Machines should take over the tasks that are dangerous, dirty, repetitive or exhausting. That frees human beings for what they do best: creating, caring, teaching, deciding and connecting with one another. The goal is to augment human capacity, not to make human beings superfluous.
A simple test can guide any company or government. Before introducing a new system, ask: does this make the people involved more capable, safer and freer, or does it make them more dependent, more fragile and easier to discard? The answer reveals whether we are optimizing for humanity or against it.
When Technology Escapes Control
The gravest danger is not that machines will rebel like the robots of science fiction. It is that a handful of people will use machines to concentrate wealth and power beyond the reach of any law, any vote and any competitor.
Advanced AI requires enormous resources: vast data centres, rare chips, huge quantities of energy and data, and specialized talent. Only a few corporations and states can afford them. This creates a natural tendency toward concentration. Whoever controls the most powerful systems can shape markets, information, and even public opinion.
In such a world, the gains from optimization flow upward. Productivity rises, but wages stagnate. Profits grow, but the workers whose jobs disappeared receive nothing. Citizens become data to be harvested, consumers to be steered, and voters to be influenced by messages tailored to their fears.
This is what I mean by a new form of slavery. It does not need chains. It works through dependence: on platforms we cannot leave, on algorithms we cannot question, on employers who can replace us at the push of a button. A person who has lost both economic independence and access to truthful information is no longer fully free.
There is also the risk of self-destruction. Autonomous weapons, AI-designed cyberattacks, systems that make decisions faster than humans can review them, and the misuse of AI in biology or chemistry all carry dangers on a civilizational scale. When competition pushes companies and nations to move faster than safety allows, the race itself becomes the threat.
Greed is not new; it is as old as humanity. What is new is the multiplier. A greedy monarch of the past could exploit a kingdom. A greedy actor in control of uncontrolled, highly capable AI could affect the entire planet. That is why the question of control is not technical alone. It is political, moral and urgent.
Lessons from History
Humanity has faced great technological upheavals before, and each time the outcome depended on whether society learned to govern the new power. History offers both warnings and reasons for hope.
The first Industrial Revolution brought steam engines, mechanical looms and an explosion of production. It also brought child labour, sixteen-hour working days and miserable slums. Between 1811 and 1816, English textile workers known as Luddites smashed machines in despair. Their violence failed, but their grievance was real: the benefits of the new machines were not being shared.
What eventually changed the situation was not the destruction of technology but its regulation. Britain’s Factory Act of 1833 began to limit child labour. Trade unions won shorter hours and safer conditions. Public schooling spread literacy. Over the following century, the same machines that had once crushed workers helped build a broad middle class.
The late nineteenth century gave another lesson. Giant trusts in oil, steel and railways concentrated economic power in a few hands. The United States answered with the Sherman Antitrust Act of 1890 and, later, the breakup of Standard Oil in 1911. Competition was restored, and innovation continued.
In the twentieth century, nuclear technology showed that some inventions are too dangerous to leave without international rules. Treaties on testing and non-proliferation did not remove the danger, but they slowed the race and kept humanity alive through the Cold War.
In 1930, the economist John Maynard Keynes predicted that technology would one day allow his grandchildren to work perhaps fifteen hours a week. Productivity did grow as he imagined, yet working hours fell far less. The reason is instructive: technology alone does not decide who benefits. People, institutions and laws decide.
The pattern is consistent. Technology creates the possibility of abundance; governance determines whether that abundance is shared or captured. Artificial intelligence will follow the same rule, only faster.
Conditions for Keeping Technology Under Control
If history teaches that governance makes the difference, then the practical question becomes: what kind of governance? I propose six conditions that together can keep optimization in the service of humanity.
- Human responsibility for critical decisions. In medicine, justice, finance, defence and education, a qualified human being must remain accountable for decisions that affect lives and rights. A machine may advise; a person must answer.
- Transparency and the right to explanation. Citizens should know when an algorithm is judging them and be able to obtain a clear explanation and an appeal. The European Union’s AI Act, adopted in 2024, moves in this direction by classifying AI uses by level of risk.
- Safety testing before deployment. We do not allow new medicines or aircraft to reach the public without rigorous testing. The most powerful AI systems deserve the same discipline, with independent audits rather than self-assessment alone.
- Fair competition and shared gains. Antitrust law must prevent a few firms from owning the entire infrastructure of intelligence. Tax systems, worker ownership and retraining programs can ensure that productivity gains reach the many, not only shareholders.
- Education for a changing world. Schools must teach not only how to use technology but how to question it: mathematics, logic, statistics, ethics and critical reading. A population that understands how algorithms work is far harder to manipulate.
- International cooperation. No single country can control a technology that crosses borders in milliseconds. Like nuclear arms control, AI safety requires agreements among nations, including rivals, on the most dangerous uses.
Businesses themselves have a role that goes beyond obeying the law. A company that invests in its workers’ skills, measures quality honestly and refuses to exploit its customers builds something more durable than quarterly profit: trust. In the long run, trust is the most profitable asset of all.
Finally, each of us carries a share of responsibility. As consumers, voters, parents and teachers, we decide which products to support, which leaders to elect and which values to pass on. Control over technology is not something we can delegate entirely to experts.
Conclusion: Technology as Servant, Not Master
To what extent can optimization be pursued for profit? The answer is: as far as it continues to respect the quality of the product, the dignity and safety of human beings, and the stability of society. Within those limits, optimization is one of the great engines of human progress. Beyond them, it becomes extraction disguised as efficiency.
Artificial intelligence and robotics could free humanity from much of its drudgery, cure diseases, reduce poverty and give every child access to an excellent tutor. They could also deepen inequality, weaken democracy, erode human skills and place unprecedented power in the hands of a few. Both futures are possible. Neither is inevitable.
The difference will be made by human choices: the laws we write, the institutions we strengthen, the education we provide and the values we refuse to abandon. A technology that moves at the speed of light requires wisdom that keeps pace with it.
As a teacher, I always told my students that a calculator is a wonderful tool, provided you still understand the mathematics. The same holds for our civilization. Let the machines calculate, sort, lift and assemble. But let human beings remain the ones who understand, who decide and who answer for the consequences. Technology must remain our servant, never our master.
Paul Sinclair, Vancouver, British Columbia
Leave a Reply