AI Assessment Tools

This category covers tools that measure people rather than models: classroom grading, game based quizzes, hiring tests, coding assessments, brain training apps and research grade cognitive batteries. They sit on one page because they share a hard question. What does a score actually tell you about a person, and how much should you trust it?

The tools fall into four groups, and the right one depends on what you need to measure.

Classroom assessment. Gradescope helps instructors grade paper, digital and code assignments against a shared rubric, with AI answer grouping on its Institutional license. Kahoot! runs live quiz games that give teachers fast formative feedback. Neither is a validated psychometric instrument; both are about scoring work fairly and quickly.

Hiring tests. Criteria and TestGorilla sell aptitude, personality and skills tests for screening candidates. Criteria works through demos and sales, while TestGorilla publishes prices and has a free tier. Short timed tests are one signal among several, so pair them with a structured interview.

Coding assessments. HackerRank and Codility test engineering candidates on realistic tasks. In 2026 both let candidates use AI inside the test and grade how well they plan, direct and review AI generated code, rather than banning it.

Cognition and brain training. CANTAB from Cambridge Cognition is a validated digital battery used in clinical research. Lumosity and BrainHQ are consumer training programs. Getting better at a training game is not the same as a measured gain in everyday cognition, and Lumosity itself says more research is needed to make that link.

A few rules apply across the whole group. Ask who validated the test and on which population. Ask whether the score is a probability, a rank or a raw count. Check whether AI is scoring people, and whether a human reviews the decision. And treat vendor case study numbers as marketing until an independent study backs them.

The same skepticism we apply to AI benchmarks applies here. Our guide to LLM benchmarks explained covers saturation, contamination and why a single headline score hides different abilities, and most of those lessons carry over to human testing.