Eight questions, derived from Magnifica Humanitas, that decide whether the AI system you are about to ship makes the world more like a Tower (concentrated power, hollowed-out humans, opaque accountability) or a rebuilt City: distributed responsibility, dignified work, plural voices, transparent governance.
"Let us build ourselves a city, and a tower with its top in the heavens, and let us make a name for ourselves." — Gen 11:4
"He convened the families, assigned each of them a section of the wall, listened to their concerns…" — Neh 3
Hold a specific AI system in mind: your own product, a vendor proposal, a competitor, a regulation. Answer each question. The verdict at the bottom updates live.
Answer all eight to receive a verdict. The Babel Test is binary by design. On each dimension, a system either builds the Tower or rebuilds the City. Drift in the middle is unexamined ground, not neutrality.
| § | Entity | Signal | Verdict |
|---|---|---|---|
| 01 | Cloudflare | 1,100 humans cut (~20% of staff). CEO confirms AI usage up 6× in 90 days. Q1 revenue record. Profit recycled out of payroll into compute. | Babel |
| 02 | Microsoft | First-ever voluntary buyout in 51-year history. Up to 7% of US workforce (~8,750). Same year: ~$80B AI capex. Big-four total: $725B. | Babel |
| 03 | OpenAI Model Spec | Document governing the behavior of a system used by 700M+ weekly users, drafted by ~200 internal staff. No external mandate, no electorate, no appeal. | Babel |
| 04 | Anthropic vs Pentagon | Refused Pentagon "all lawful purposes" clause. Held red lines on lethal autonomous weapons and domestic surveillance. Trump's response: "fired like dogs." | Jerusalem |
| 05 | EU AI Act · Impact Assessments | Mandatory fundamental-rights impact assessments. Procedural, distributed, transparent. Slower deployments. Many voices, common ground. | Jerusalem |
| 06 | Anthropic on the Papal Dais | Co-founder shares stage at Vatican Synod Hall. Concedes incentives "can sometimes conflict with doing the right thing." Vatican source: "not an endorsement." | Mixed |
"The main drivers of development are private, often transnational, parties that are endowed with resources and the capacity to intervene that surpass those of many Governments."
"Technology is never neutral, because it takes on the characteristics of those who devise, finance, regulate and use it."
"The 'Babel syndrome' — the idolatry of profit that sacrifices the weak, a uniformity that neutralizes differences, and the pretense that a single language, even a digital one, can translate everything, including the mystery of the person, into data and performance."
"Ownership of data cannot be left solely in private hands but must be appropriately regulated… patents, algorithms, digital platforms, technological infrastructure and data."
"A society that guarantees employment to only a small fraction of the population, despite having a high level of technical development, risks… a paradox of material progress and anthropological regression that undermines the foundations of a just and stable social peace."
"Disarming AI means freeing it from the mentality of 'armed' competition, which today is not limited simply to the military context, but is also an economic and cognitive phenomenon."
The Babel Test is a distillation. Eight questions, drawn from the structure of Pope Leo XIV's Magnifica Humanitas, framed so an engineer can run them on a deployment review on Monday morning.
The Babel Test is the governance heuristic the AI discourse has been refusing to write. The people writing AI governance work for the labs they should be scoring.