A recent study found that experienced software developers using AI tools to complete tasks took 20% longer than when they didn't use the technology, challenging the assumption that AI always boosts workplace efficiency. The study, conducted by Model Evaluation and Threat Research, tasked 16 software developers with completing work using AI tools.

The researchers found that the developers, who initially expected AI to expedite their work, experienced a decrease in productivity. This outcome contradicts the prevailing narrative of AI driving significant gains in workplace efficiency.

The study highlights a potential disconnect between the promise of AI and its actual impact on software development workflows. While AI tools are often marketed as productivity enhancers, this research suggests that their integration may not always be seamless or beneficial. The specific AI tools used in the experiment were not disclosed.

The findings raise questions about the current state of AI in software development and the need for further research into how these tools can be effectively integrated to improve, rather than hinder, developer productivity. It also suggests that developers may need more training and experience to effectively leverage AI tools.

The Model Evaluation and Threat Research group plans to continue investigating the impact of AI on various industries, focusing on identifying best practices for AI implementation and mitigating potential drawbacks. Future studies may explore the impact of different AI tools, training methods, and task complexities on developer productivity.