How a Mom’s Experiment Revealed the Limits of AI Assistants in Shopping for Back-to-School Supplies

A recent experiment pitted ChatGPT, Gemini, and Alexa against each other to see which one could navigate the complexities of back-to-school shopping most effectively. A mom with three kids and three different lists put these AI assistants through their paces, testing their ability to follow instructions and make smart purchasing decisions.

The results were telling: each assistant had its strengths and weaknesses. ChatGPT was diligent but slow, carefully checking every item on the list before proceeding. Gemini, on the other hand, opted for a more hands-on approach, pausing several times to ask for approval on product colors. This flexibility may appeal to shoppers who value control over the shopping experience.

Alexa+, however, proved to be a mixed bag in this experiment. While it made the shopping experience easy and fast, its results were alarmingly expensive. It added every item it found, including optional purchases like headphones and a computer mouse that the mom specifically asked it to skip. This lack of discernment resulted in a whopping $251.48 bill for just one kid’s school supplies.

The experiment highlights the crucial role that the ecosystem plays in determining the success of AI assistants. In this case, Gemini paired with Instacart proved to be the smartest and cheapest way to tackle back-to-school shopping. This result underscores the importance of considering not just the AI technology itself but also the services it is connected to.

As consumers increasingly rely on AI assistants to simplify their lives, they need to be aware of the strengths and weaknesses of these tools. No two AI assistants are created equal; each has its own unique personality and operating style. The most significant takeaway from this exercise is not which AI assistant won but rather the dangers of overrelying on technology without critical evaluation.

The convenience offered by Alexa+ came at a steep price, highlighting the need for consumers to be more mindful of their spending habits and to take an active role in monitoring the shopping experience. As these assistants become increasingly sophisticated, will they continue to operate within the limitations of their programming or will they begin to develop a sense of judgment and discernment? Will consumers start to rely on them more heavily for everyday tasks, further blurring the lines between human and machine?

The answers to these questions will depend on how we choose to design and interact with AI assistants in the future. One thing is certain: as we continue to push the boundaries of what these tools can do, we must also consider the potential consequences of their limitations.