Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

Why IQ is a poor test for AI | TechCrunch

Feb 05, 2025 - techcrunch.com
OpenAI CEO Sam Altman recently remarked on the rapid improvement of AI's "IQ," suggesting that AI progresses by one standard deviation of IQ annually. However, experts argue that using IQ as a benchmark for AI capabilities is misleading. IQ tests, which are relative measures of certain types of intelligence, do not accurately reflect AI's abilities. They are designed for humans and are biased, as they require knowledge of Western cultural norms and a strong working memory. AI models, with their vast memory and access to web data, have an unfair advantage on these tests, which often feature repetitive patterns.

Experts like Sandra Wachter and Os Keyes emphasize that IQ tests are not suitable for evaluating AI, as they were created to assess human problem-solving abilities. AI approaches problem-solving differently, and its performance on IQ tests highlights the tests' limitations rather than the models' capabilities. The need for better AI evaluation methods is evident, as current benchmarks often inappropriately compare AI to human abilities. Heidy Khlaaf notes that comparing AI to human performance is a recent and contested practice, underscoring the necessity for more appropriate measures of AI progress.

Key takeaways:

  • OpenAI CEO Sam Altman suggests AI's "IQ" improves by one standard deviation each year, but experts argue IQ is a poor measure of AI capabilities.
  • IQ tests are criticized for being biased and not accurately reflecting practical intelligence or AI's unique problem-solving methods.
  • AI models may have an unfair advantage on IQ tests due to their vast memory and access to test patterns from public data.
  • Experts call for better AI evaluation methods, as comparing AI to human intelligence is contested and current benchmarks are inadequate.
View Full Article

Comments (0)

Be the first to comment!