“Come, let us build ourselves a city, with a tower that reaches to the heavens, so that we may make a name for ourselves; otherwise we will be scattered over the face of the whole earth.” Genesis 11:4

The Tower of Babel story is only a handful of verses, but it moves fast. People unite, make a plan, start building, and then the whole thing collapses.

In a project the size and scope of the Tower, it would not be unreasonable to assume one of  the biggest challenges facing the developers was disagreement, with people having different opinions, different goals, different languages. But in Babel, the surprising thing was that everyone was aligned. They were all on the same page, speaking the same language, moving with the same energy and purpose. It was perfect teamwork.

And yet, it still goes wrong, because danger wasn’t only in the differences. The story suggests that total unity became its own risk when it was paired with pride, unchecked ambition, and a sense of invincibility. In other words, Babel isn’t a cautionary tale about miscommunication. It’s a cautionary tale about overconfident coordination. Even when humans finally get perfect collaboration, it doesn’t automatically lead to wisdom or good outcomes.

The Tower of Babel becomes a monument to the fact that humans are never more dangerous than when they are perfectly aligned and just slightly too proud of it. In today’s tech era, we might be rebuilding Babel, only this time it is made of servers, APIs, single sign-on, and a Terms of Service agreements (no one reads). It is not a tower, technically, but a stack. And it is getting taller every day. Babel in the Cloud.

The key ingredient in the Babel story is not bricks but standardization. One language meant everyone could communicate instantly, coordinate effectively, move in the same direction, and scale a plan without friction. That is also the dream of modern technology, especially the kind of technology that AI rides on. When people talk about global-scale models and shared infrastructure, what they really mean is that we are building systems that speak one language, run on the same few clouds, authenticate through the same identity providers, and increasingly funnel knowledge through the same interfaces. The more it works, the more everyone adopts it. The more everyone adopts it, the more essential it becomes. And the more essential it becomes, the closer we get to a single point of failure.

The Babel story almost has an entrepreneurial optimism to it, the kind that would look right at home in a pitch meeting where everyone is wearing matching hoodies and saying the word “scale” like it’s a mantra. Genesis 11:16, “And the LORD said, Behold, the people is one, and they have all one language; and this they begin to do: and now nothing will be restrained from them, which they have imagined to do,” hints at the mindset at many technology companies today, that nothing they plan to do will be impossible for them. That sentence could be printed on the wall of a dozen companies, and nobody would blink.

This is where the Babel story becomes an allegory for the AI era, because modern AI is not just a tool, it is a multiplier of speed, coordination, and ambition. It inspires awe because it can write, summarize, tutor, code, draft legal language, and imitate a tone that sounds suspiciously like a competent adult. It creates acceleration because tasks that once took teams and weeks can be done by one person in an afternoon. It creates competitive pressure because nobody can afford to be the only shop in town that still does things the old-fashioned way. And it creates what might be called inevitable-ism, the sense that whether you love it or fear it, you might as well get on board because the train has left the station and the conductor is an algorithm.

The danger is not that this energy exists but that it tempts people toward moral shortcuts. Overconfidence and centralization are an especially flammable combination. The we-can-do  anything spirit is thrilling until it becomes the justification for going faster than consequences can be forecasted and guardrails can be placed. If Babel had a pitch deck, the Risks & Mitigations slide  would contain the letters “TBD” in a font large enough to be seen from the heavens.

The modern caution embedded in Babel can be summarized by noting that when everyone builds on one stack, errors and abuses scale fast. Centralization is not automatically bad, because standardization is how we get things like reliable internet protocols, electrical grids, and medical guidelines that save lives. The problem is that the more centralized the system becomes, the more catastrophic the failure can be. A broken toaster ruins breakfast, a broken grid ruins a city. And AI, especially AI embedded into institutions, has the potential to become grid-level important. That means the failure modes are potentially catastrophic.

One plausible near-future version of Babel is not a literal tower but a default platform. Imagine a single AI system becoming the standard tool for education, where it tutors students, grades assignments, recommends learning paths, and shapes admissions decisions. Imagine that same system becoming standard in hiring, where it screens resumes, generates interview questions, scores candidates, and predicts culture fit (a concept that has never been misused in the history of mankind and is surely safe to automate). Imagine it becoming standard in policing and risk assessment, where it helps determine who gets flagged, monitored, prioritized for intervention, or routed into different outcomes. Imagine it becoming standard in finance, where it drives credit decisions, fraud detection, loan approvals, and transaction monitoring. Now imagine it contains a subtle flaw born from training data that reflects old prejudices against, just say, the Jews. In a decentralized world, such errors appear in pockets. In a Babel world, they become society-wide.

There is another Babel-like risk that is quieter but just as important, homogenization. When everyone uses the same AI to write, you can feel a sameness creeping in. Websites begin to share the same confident-yet-friendly tone and essays read like they were written by the same diligent person who is uncomfortable with humor. As annoying as it is, this is not the main danger. The bigger problem emerges when everyone uses the same AI to make decisions, because then the same blind spots repeat, the same assumptions harden into an objective truth, the same mistakes become policy, and the same invisible ideology gets mistaken for neutrality. One language is not only about communication, but also about worldview. When a single worldview becomes dominant enough it stops being questioned, and AI is exceptionally good at sounding unquestionable.

An even scarier version of a failure is one that is smooth, where an update quietly changes thresholds, misclassifies people, flags harmless behavior as suspicious, or shifts priorities in ways no one notices until the damage accumulates. Centralized systems do not always break with a bang. It’s possible to break with a virtual smile.

This is where Babel becomes moral philosophy rather than infrastructure analysis. When systems become powerful and widely shared, people start treating them like law. Accountability gets lost to the algorithm. In the Babel story, the humans are trying to escape limits. They want to reach the heavens. They want to make a name and to be untouchable. In the AI era, the temptation is to build systems so efficient they feel above accountability, to automate decisions so no one has to own them, and to claim neutrality because the model is supposedly objective. But morality does not vanish when you put it into a workflow, it only becomes harder, because harm becomes indirect, distributed, and deniable. Deniable harm is the kind humans are most tempted to tolerate.

So what does an anti-Babel strategy look like, if we want to keep the benefits of coordination without creating a fragile monoculture? It begins with remembering that Babel is not anti-technology but anti-hubris. A practical response is to avoid monocultures where possible, because diversity reduces correlated failure. If institutions rely on multiple models, multiple vendors, and multiple evaluation methods, they reduce the risk that one blind spot becomes universal. It also means building human accountability into the loop in high-stakes settings so that the model never becomes a moral escape hatch. Decisions that shape lives need transparent criteria, oversight, and appeal processes, because people deserve a way to contest mistakes that look official.

An anti-Babel approach also treats auditing as normal rather than scandal-driven. Bias testing, red-teaming, and post-deployment monitoring should be routine practices, not rituals performed after a headline. It draws a bright line between assistance and decision-making, because AI can be excellent at summarizing, suggesting, and supporting while still being unreliable as a judge of character or context. It also plans for failure as seriously as it plans for success, because a system that cannot fail gracefully is a system that will eventually fail catastrophically.

Every AI integration should come with what can be called The Babel Warning label: This tool increases speed, scale, and coordination, and may also increase systemic risk, centralized failure, unaccountable decision-making, overconfidence, and moral shortcuts. Users are advised to proceed with humility and to avoid attempting to reach the heavens without a backup plan.

The surprising gift of the Babel story is that it can be read not only as punishment but as protection. A world with one language and one unified project can become a world with no friction, no dissent, no local variation, and no meaningful resistance to bad ideas. Fragmentation is inconvenient, but it is also a safety feature. Different languages force translation, which forces thought, which in turn forces humility. And humility, inconveniently for all of us, is the one thing AI cannot automate.

So yes, we should build powerful tools. We should use AI, innovate, and enjoy the productivity gains. We must however be diligent to never confuse coordination with righteousness, scale with wisdom, or build one tower so tall that when it wobbles it takes everyone with it.

© The Times of Israel (Blogs)