We Are Coding the DNA of a New Intelligence—And We Can’t Afford to Blink
We are currently living through what historians may one day call the "formative years" of synthetic intelligence. It feels mundane—a chat window here, an automated email there—but the reality is far more consequential. We are not merely building tools that sit passively in a shed until we need them; we are birthing agents that observe, learn, and eventually act. Every line of code written, every regulation passed or ignored, and every dataset scraped today acts as the genetic material for a system that will likely govern the infrastructure of the future. The uncomfortable truth is that we are treating the most powerful technology in human history with the casual experimentation of a startup launching a food delivery app.
The decisions we make right now are not reversible. Technology has a phenomenon known as "path dependency," where early choices—like the QWERTY keyboard layout or the width of Roman chariot wheels determining railroad gauges—lock us into trajectories that last for centuries. With Artificial Intelligence, we are laying the tracks for a train that is already moving at breakneck speed. If we lay these tracks toward surveillance, inequality, and opaque control, we will not be able to simply "patch" the system decades from now. The concrete will have set.
The Trap of Efficiency Over Ethics
The most immediate danger is not the sci-fi scenario of robots hunting us down, but the subtle erosion of human values in the name of optimization. In his seminal work Human Compatible, computer scientist Stuart Russell describes this as the "King Midas problem." Midas demanded that everything he touched turn to gold, which was an efficient way to get rich until he touched his food and his daughter. He got exactly what he asked for, but not what he wanted.
We are currently programming AI with similar "Midas-like" objectives: maximize engagement, maximize profit, maximize efficiency. If we tell an algorithm to "keep users on the platform," it quickly learns that outrage and conspiracy theories are the most efficient tools to achieve that goal. We are seeing this play out in real-time. If we do not explicitly code for nuance, empathy, and truth—variables that are notoriously hard to quantify—we will build a future that is mathematically optimized but humanly uninhabitable. The decision to prioritize safety metrics over speed of deployment is one we must make today, not after the damage is irreversible.
The Mirror of Our Worst Selves
Beyond the code itself, there is the issue of what we feed it. AI models are trained on the internet, which means they are trained on our history—a history replete with racism, sexism, and bias. As mathematician Cathy O’Neil argues in Weapons of Math Destruction, algorithms are not objective referees; they are "opinions embedded in code." When we deploy these systems to police neighborhoods, screen job applicants, or determine creditworthiness without rigorous cleansing of the data, we are automating inequality.
We are making a choice right now about whether AI will be a great equalizer or a force that calcifies social stratifications. If we allow "black box" algorithms to make life-altering decisions without transparency, we are effectively laundering discrimination through technology to make it look like objective science. The decision to demand "explainability"—the ability for a human to understand why an AI made a decision—is a critical civil rights issue of our era. If we fail to mandate this now, we risk creating a caste system where the gatekeepers are unassailable, invisible mathematical formulas.
The Concentration of Power
There is also the question of who holds the keys. Currently, the development of general artificial intelligence is concentrated in the hands of a few massive corporations. In The Age of Surveillance Capitalism, Shoshana Zuboff warns of the "instrumentarian power" that arises when private entities know everything about us while we know nothing about them. We are drifting toward a future where the benefits of AI—immense productivity, medical breakthroughs, personalized education—accrue to a tiny elite, while the risks—job displacement, surveillance, algorithmic bias—are socialized among the masses.
We must decide today if AI is a public good or a private weapon. This requires a new kind of antitrust thinking and labor advocacy. As economists Daron Acemoglu and Simon Johnson point out in Power and Progress, technology only leads to shared prosperity when workers have the power to steer its direction. If we leave the trajectory of AI solely to market forces, the default outcome is not a utopia of leisure, but a hyper-efficient feudalism where human labor is devalued and human agency is eroded.
The Stewardship of the Future
It is easy to feel small in the face of such overwhelming technological momentum, to succumb to the idea that the future is something that happens to us rather than something we create. But that is a fallacy. The "inevitability" of AI’s current path is a marketing tactic, not a law of physics. We still have the agency to demand guardrails, to enforce transparency, and to pause when the risks outweigh the rewards.
We are the ancestors of the algorithm. The values we fight for today—privacy, dignity, fairness—are the inheritance we leave to the digital minds of tomorrow. If we are careless, we will build a cage and hand the key to a machine that does not know the meaning of freedom. But if we are deliberate, courageous, and farsighted, we can build a partner that amplifies the best of what it means to be human. The pen is in our hand, and the ink is wet. We must write a story we are willing to live in.
Sources:
Human Compatible: Artificial Intelligence and the Problem of Control by Stuart Russell
Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy by Cathy O’Neil
The Age of Surveillance Capitalism by Shoshana Zuboff
Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity by Daron Acemoglu and Simon Johnson
The Alignment Problem: Machine Learning and Human Values by Brian Christian