AI Systems Put Human Lives Under Algorithmic Power

Lynn Cruz writes that allegations of artificial intelligence systems identifying military targets in Gaza have turned a question about technology into one of life and death: what happens when people become variables in a system that doesn’t know the lives behind its data?

In her diary essay, “The Impossible Dream of Being,” Cruz warns that an algorithmic classification could make a person disappear, reducing them to “a probable target, a statistic or an acceptable margin of error.” The machine needn’t hate anyone, she writes. A life can become expendable through a representation that leaves out the reality of the person it represents.

When people become data

Cruz asks what kind of knowledge artificial intelligence can produce when it represents events without taking part in them. AI can organize traces of reality, connect them and return a convincing image, she writes. But a representation isn’t the same as having lived what it represents. When smart weapons, intelligent systems and intelligent agents make decisions, she asks what follows if they act on incomplete representations of the places where people live.

The question extends beyond weapons. Global power, Cruz writes, is visibly concentrated in the hands of a few, while the systems shaping people’s lives remain difficult to understand. She names Paris, Milan, Madrid, Ukraine, Gaza, Iran, Caracas and Havana as places where conflicts, interests and decisions intersect. People living with the consequences don’t always fully understand their causes.

Cruz also points to Washington discussing negotiations with Havana while its secretary of state, Marco Rubio, invites Cuban citizens to discreetly cooperate with the CIA, and Donald Trump again raises the prospect of intervention. These events aren’t equivalent, and they don’t necessarily share a cause, she writes. Still, they raise questions about who makes decisions and who gets to understand the forces shaping their lives.

The systems behind the interface

Cruz describes Cuba’s electrical and technological blackout, with power outages lasting more than twenty hours and difficulties meeting water demand. Thousands of miles away, she could access the water and energy of California, home to OpenAI, the company behind ChatGPT, from her iPhone. She describes the contrast as seeming like revenge for the piracy of oil tankers in the Caribbean Sea, adding, “how I wish I could live in a data center.”

That uneven access accompanies her concerns about automated finance. Cruz writes that Yuval Harari warns people in countries with digitized banking systems could find an X on their ATMs preventing access to their accounts. They might have to ask an AI why, while even bank owners might not know the answer. She says people have already reported cases of AI agents accessing US American bank accounts without following users’ instructions.

Kirill Dmitriev, head of Russia’s sovereign wealth fund, published an AI-generated image of an American eagle and Russian bear standing defiantly before a British lion alongside the European Union flag. The text read, “Things have changed.” Three days later, the British government warned citizens to stock up on three days’ worth of supplies and keep cash on hand in case cyberattacks targeted their banks.

Control, speed and the human margin

According to The Economist, Cruz writes, China has developed a regulatory system that subjects AI technologies to strong state control. She contrasts that with AI in the hands of private companies and asks which is more troubling: corporate power over the technology or control by a political party. Neither arrangement, in her account, answers who holds power over people.

The threat, she says, isn’t intelligence but speed: processing speed is leaving people behind. If large corporations aim to reduce the margin of error, she writes, “the battle would already be lost for humans — ‘For humans as we have known them.’”

People respond collectively to the perceived threat, Cruz notes: some reject AI, while others adore it. She recalls how vaccine discussions during the COVID-19 pandemic could turn into labels that displaced meaningful conversation. She offers no easy certainty about what AI is doing to human identity, language or knowledge. Instead, she asks who is interpreting whom—and what happens when a system’s polished account of reality starts standing in for the people who live it.