As AI companies compete to develop digital assistants capable of handling complex tasks like travel planning and shopping, a recent experiment tested the current utility of these tools. The journalist utilized the WNBA Finals as a practical scenario to evaluate the performance of OpenAI's 'Dots' and Meta's 'Muse.'

The core objective was to determine if these AI agents could meet a 'Human Benchmark,' defined as the ability to execute tasks as quickly and accurately as a human counterpart. While the experiment highlighted the rapid development of these technologies, the initial results suggest that these AI agents are not yet fully capable of replacing human decision-making or efficiency in travel planning. The journalist noted that while the tools are advancing, there remains a significant gap between the current functionality of these AI agents and the level of trust required for users to delegate significant personal tasks, such as managing finances or complex travel itineraries.