Dropping the concept of race: AI sidesteps the science

Human genomes are 99.9% identical, yet AI refuses to classify human races while happily detailing 364 dog breeds.

English · Original discussion in Spanish · Published

AI denies human races but classifies 364 dog breeds: the uncomfortable science

Ask an artificial intelligence about physical differences between people of different races and you will get evasions, caveats and vagueness. Ask it about dog breeds and it will answer with a precise taxonomy: the Fédération Cynologique Internationale recognizes 364 varieties. That asymmetry condenses an uncomfortable dilemma that modern genetics resolves with one figure: the genome of any two humans is 99.9% identical.

Biology versus statistics

Science holds that race has no biological validity when applied to humans. But the uncomfortable statistics insist that measurable differences exist in height, eye tonalidad or skin. Those who defend the concept point to selective breeding: if we distinguish 350 breeds of pigeons, why not in the most diversified species on the planet? The standard answer is that human variation is continuous, not discrete.

Race, ethnicity and power structures

Some propose replacing race with ethnicity. The nuance is not merely semantic: China is organized around the Han ethnicity, with 1.4 billion people, rather than 28 countries of 50 million. The cohesion of interethnic authority explains many conflicts. There, the concept of race matters, whether we like it or not.

The problem is that abandoning the word does not eliminate the data. It only makes them harder to say out loud.

Also available in: Deutsch

Summary of a discussion on Burbuja.info - Foro de economía, actualidad y política., translated from Spanish and reviewed before publication. Read the full discussion (11 replies).

More summaries

All summaries in English →

-
Back