In this project I explored how to support creative teams in the advertising industry who face intense pressure to generate numerous ideas under tight deadlines. I designed and developed "Phantom," an AI-powered text editor that functions as a collaborative partner in the creative process. The tool uses an LSTM-RNN machine learning model to provide text-based suggestions, aiming to inspire users and increase their creative output. To validate this approach, I conducted an experiment with 30 advertising professionals divided into three groups: one using the AI, one with no tool, and one with a "dummy" system providing random text suggestions. The results showed that collaborative tools significantly increased the quantity of creative output. Interestingly, while the AI-generated text was perceived as more relevant, the group using the dummy system produced work that was rated highest in quality by a panel of independent judges. This research highlights the potential of human-AI collaboration in ideation and raises new questions about the nature of creative inspiration.

Read here for more detailed documentation about the project.