Machine Intelligence Test: How AI Is Measured
From the Turing Test to AI scoring 140+ on Mensa — how machine intelligence testing actually works, and why the headline numbers mislead.
What IQ is, how IQ tests work, the history and science of intelligence measurement.
28 articles
From the Turing Test to AI scoring 140+ on Mensa — how machine intelligence testing actually works, and why the headline numbers mislead.
Learning styles, the 10% brain, left- vs right-brained learners — surveyed teachers endorse these at strikingly high rates. What the evidence says.
Every cognitive test correlates with every other. g is the model extracted from that pattern — not a substance in the brain. What the evidence shows.
The Wechsler Adult Intelligence Scale is the world's most used adult IQ test. Its editions, index scores, the General Ability Index — and why independent researchers reject the WAIS-5's five-factor structure.
The Cognitive Reflection Test answers are 5 cents, 5 minutes and 47 days. Why the intuitive answers fail, what your score means, and whether knowing the answers really invalidates it.
Cognitive tests fall into three unrelated categories. How clinical screens like the MoCA and MMSE differ from IQ tests, what the scores mean, and why the 26/30 cutoff misleads.
Ability tests, trait questionnaires and mixed models measure different things. How the MSCEIT is scored, why self-report EI is mostly personality, and why there is no EQ score.
There are no left-brained or right-brained people — 1,011 brains and 14 million connections found none. What is genuinely lateralised, and what Sperry warned about.
The SB5's five factors, its adaptive routing, and the standard-deviation change from 16 to 15 that makes old Stanford–Binet scores incomparable with new ones.