Chatbots: Uncanny Valley or Your New Best Friend?

Have you ever experienced an overwhelming sense of discomfort when being faced with a photo or video of a robot or toy resembling a human, despite there being no obvious danger to you? Well, you’re not alone. This phenomenon — yes, it’s a studied phenomenon — is called Uncanny Valley, and refers to the correlation between the human-like appearance of an object and the significance of the emotional response endured by the viewer.

Originally introduced by Tokyo Institute of Technology professor Masahiro Mori, Uncanny Valley states that favorable feelings towards robots increase as the robot’s human likeness increases, but with limits (Caballar 2019). The “valley” refers to the drop in favorability after human likeness reaches a point after 50%, as illustrated in the following graph created by Mori. It is important to note that after the dip of the valley, the favorable feelings rise once more as the object approaches 100% human resemblance, “indicating a living person.” (Kendall).

A resemblance to human behavior is a fundamental part of artificial intelligence in our daily lives. One example of this is the presence of chatbots — computer-simulated conversational agents — often used on e-commerce websites to improve the speed and quality of live customer service. According to Forbes, there are two main types of chatbots: declarative, which operate from a bank of scripted responses and are often used on company websites, and predictive, best described as digital assistants like Siri and Alexa (Main and Schweta 2022). Chatbots are an efficient way to cut labor costs while increasing the availability of customer support. Because chatbots are built to recreate human interaction, they’re at risk of falling into the Uncanny Valley. But, because chatbots do not typically have a physical form resembling a human, they’re less likely to spur those significant feelings of discomfort often associated with Uncanny Valley.

One example of a company using a chatbot in a unique and meaningful way is Woebot, a virtual assistant that helps you talk through mental health struggles from the privacy and comfort of your home. According to the New York Times, Woebot “uses artificial intelligence to deploy the principles of cognitive behavioral therapy.” (Brown 2021). Woebot is embarking on a journey taken by many companies developing AI solutions — how can typical human processes be made to be more efficient, more affordable, and offer greater value to the consumer?

For one, Woebot offers users the opportunity to talk about their mental health at any hour of the day, and have something respond to them. Woebot also presents a far more affordable option than traditional therapy. Although there are certainly concerns about the effectiveness of therapy chatbots like Woebot (can a declarative bot really stop someone from harming themselves or others?), one thing is for certain: chatbots offer a new level of human effectiveness that is actively changing the world as we know it.