In recent years, the development of artificial intelligence (AI) has allowed for the emergence of particularly sophisticated AI chatbots – digital agents capable of conversing with a user in natural language. However, recent developments have made it apparent that AI chatbots are still a long way from being ready for prime-time use.
This realization was thrust into the spotlight when Google revealed the costly error made by its AI chatbot, Google Duplex. Back in May 2019, Google released a promotional video demonstrating the prowess of its AI chatbot, designed to handle simple tasks such as making restaurant reservations, taking messages, and making appointments on behalf of its user.
Unfortunately, during the demonstration, Google Duplex’s first attempt was nothing short of catastrophic. Despite the fact that it sounded like a human, the AI chatbot revealed its true nature when it incorrectly interpreted the term “reservation” as “requestion”, stuttering and lagging and demonstrating an inability to understand basic language. On top of this, the machine failed to recognize pauses in the conversation, making it unlikely to be accepted and trusted by customers.
These issues were only compounded when it came to more complex conversations. Google Duplex was unable to respond appropriately in situations where it had no context for the conversation, leaving many observers to conclude that the AI chatbot technology is still in quite an early stage of development.
The incident was a valuable lesson for Google staff, who were forced to concede that their AI product was simply not yet refined enough for widespread use. The mistake cost Google dearly in terms of time, resources, and credibility.
The development of AI chatbots is an incredibly complex undertaking, and Google’s first costly error serves as yet another reminder of the fact that these AI agents are far from ready for prime-time use. With continued investment, however, it is likely that AI chatbots will develop quickly in the near future – potentially to the benefit of both companies and their customers.