Have you been down the Uncanny Valley?
Our interactions with bots have increased tenfold during the pandemic but most of the time we don’t even realise it
In the beginning of artificial intelligence (AI) development the “Uncanny Valley” was a theoretical state of mind. Professor Masahiro Mori of the Tokyo Institute of Technology postulated that humanity should not aim to perfectly emulate humanness in the pursuit of AI and robotics.
Why?
Because when people come to encounter AI that they believe to be human but is eventually revealed to be fake, there is an inevitable feeling of shame, horror and distrust. This, Mori wrote, would cause big problems for human-bot relations.
So far, so Bladerunner.
But as we enter a new age of chatbot interaction, should the Uncanny Valley not be higher on the agenda for bot-builders?
Chatbots: winners of the pandemic
Chatbots can easily be defined as “software applications capable of communicating through natural language processing”. In short, a chatbot is trained through Machine Learning to respond to a human interlocker on the other end of a platform. Technically, the more conversation a chatbot has, the smarter (and more “human”) it will become.
The annual chatbot report from Tidio tells the extraordinary rise of chatbots in 2020.
In 2019, the global chatbot market size was $2.6 billion. In 2020, that mushroomed to $3.4 billion, a staggering 32.5% increase.
It's impossible to gauge how much of that growth was expected and how much was directly related to the global lockdown, but it’s fair to say that chatbots have been some of the real winners of the pandemic.
As chatbots have clearly proliferated, isn’t it time to ask yourself, when was the last time you spoke to a robot?
Or, more precisely, when was the last time you spoke to a robot that you thought was a human?
When was the last time you spoke to a robot?
Most people will have experienced a bot as they shop online. Bots welcome you as you land on the site and operate to convert your interest into sales. For example, if you have placed something in your virtual basket, the bot will remind you to check out (or hopefully do something savvy like offer you a promo code).
Chatbots are also increasingly appearing in more novel and surprising settings.
Duolingo and Mondly have multilingual chatbots to help users learn new languages, whilst therapist chatbots like Woebot and Wysa are available around the clock to respond to people with mental health concerns.
Despite these admirable use cases, bots are often presented with a human avatar: a photo of an attractive customer service agent called something likeable like “Jenny” or “Benji”.
But isn’t this attempt to deceive the very danger that Prof Mori warned about 50 years ago?
And if so, why are designers ignoring the real fear of the Uncanny Valley today?
If you're interested in hearing more about chatbots, BotSupply.ai CEO and founder Francesco Stasi recently appeared on Everyday Startups podcast with Dan and Nana Parry. Available on Apple & Spotify
