Large language models are proficient in solving and creating emotional intelligence tests

2026

Open publication workspace · Sign in to read the full PDF.

AI-generated summary

1) Large language models demonstrate proficiency in solving and creating emotional intelligence tests, outperforming humans and showing potential for emotionally intelligent AI applications.
2) * LLMs significantly outperform human performance on five standard emotional intelligence tests.
* ChatGPT-4 can generate new test items that are statistically equivalent in difficulty to original tests.
* While some differences exist in perceived clarity and realism, the generated tests are strongly correlated with original ones.
3) Large Language Models, Emotional Intelligence, AI, ChatGPT, Cognitive Empathy

Check the original publication for accuracy and context.