A recent study by researchers from Brown University and Chinese universities demonstrates that AI chatbots like OpenAI’s ChatGPT can effectively operate a software company with minimal human intervention. In this experiment, AI bots powered by ChatGPT’s 3.5 model were assigned roles in a hypothetical software development company called ChatDev, following the waterfall model stages: designing, coding, testing, and documenting.
The AI bots interacted and collaborated autonomously, making decisions and completing various software development tasks with minimal human input. The study found that ChatDev could complete the software development process in under seven minutes at a cost of less than one dollar on average, with around 86.66% of generated software systems executed flawlessly.
While the study identified limitations, such as errors and biases in language models, it suggests that AI chatbots have the potential to significantly streamline and cost-effectively handle certain job functions, including software development. This finding highlights the versatility of generative AI technologies like ChatGPT in various industries and their potential to assist junior programmers and engineers in real-world scenarios.