Case Study: Google DeepMind improves comedy research recruitment and workshop execution with Prolific

Google DeepMind uses Prolific to score 54.6/100 for LLM comedy writing

Google DeepMind researchers aimed to understand the effectiveness of large language models (LLMs) as creativity support tools for comedy writing. Their challenge was finding professional comedians experienced with AI and creating an environment for an authentic evaluation of AI-generated humor.

Prolific provided the solution by recruiting specialized comedy professionals and seamlessly organizing a series of workshops. This allowed researchers to gather rich qualitative and quantitative data. The results revealed that comedians found LLM-generated content bland and lacking essential human qualities, culminating in a mediocre Creativity Support Index score of 54.6 out of 100. Prolific enabled this pioneering research into human-AI collaboration.

Google DeepMind