Two humans, one computer. Almost 60 years ago, computer scientists Alan Turing and Joseph Weizenbaum envisioned a future where computers would be able to express a level of intellect that would render them indistinguishable from humans.
This question gave rise to the Turing Test: a trial-by-fire where a computer and a human are asked various questions while a judge attempts to distinguish between the two.
Eliza, the first chatbot, was developed in 1966 and used just 200 lines of code. A chatbot is basic form of artificial intelligence software that can answer questions, converse and assist us in a way that mimics humanlike conversation. Forty years later, a chatbot named Eugene Goostman successfully convinced a panel of judges that he was a 13-year-old boy, thus beating the Turing Test. The Turing Test was a turning point, with chatbots now being used in social messaging, e-commerce and enterprise platforms, and healthcare. At first, chatbots will only perform simple administrative tasks such as scheduling appointments or answering simple questions, but soon they will be able to correctly diagnose, refer and eventually treat everything from mental disorders to physical illnesses. Think an empathetic, supportive and less romantically inclined version of Samantha.
The app is dead, long live the app
The dog days of the individual app are drawing to a close. For the past eight years in the virtual wild west land-grab that is Silicon Valley, fortunes have been staked on building stand-alone applications. The victorious conquerors are messaging applications, beating out social media applications for the first time with more monthly active users in the first quarter of 2015. But, these applications stand divided, unable to communicate with each-other. The future lies in consistent cross-platform collaboration, and the leader of the next virtual revolution will be the humble chatbot. Cross-platform compatible chatbots will act as facilitators, allowing users to initiate conversations and coordinate activities through messaging in what were formerly standalone applications.
The healthcare industry has not been sidelined in the spirited arms race to build standalone apps, with many companies building their own healthcare applications for everything from medication adherence, substance cessation, self-reporting and monitoring of symptoms, and mindfulness coaching. With mixed results, the mediocre performance of mobile health applications could be in part due to the chore of interacting with a stand-alone application which reminds people of their own shortcomings and illness. With the number of chat applications users worldwide predicted to reach 2 billion by 2018, or 80% of smartphone users, the voice of reason suggests that we follow the breadcrumbs to the user. It stands to be self-evident that interacting with users directly inside chat interfaces is far more effective than directing users to standalone applications. Building healthcare chatbots into messaging applications will reduce the healthcare economic headwinds we face and change the way that we interact with our own health, creating a contextual conversation around users.
A sympathetic synthetic symphony
The human face of interactions may have passed its zenith as we increasingly turn to our devices for information, companionship and guidance. But rather than replacing emotional connection with callous transactions, chatbots could signal the rise of a conversation that eventually leads to human-to-human interaction. Joy, a chatbot developed on Facebook’s Messenger platform that tracks users mental health state-of-mind, does not seek to replace therapists, but acts as a gateway chatbot to encourage a conversation that eventually culminates with the user seeking qualified help if required.
As humans, we are disadvantaged by our physical and emotional unavailability. While social media has created a convoluted web of connection, it has also led to insincere interactions which have ironically created superficial walls. Backed by machine learning, chatbots are always on, always there, and always able to remember past conversations while learning from new ones. Chatbots are becoming more human-like, and as we continue to experiment with sympathetic chatbots, soon we will find ourselves telling our most intimate thoughts and concerns to a digital imitation of a friendly face. Microsoft has already created a sympathetic synthetic ear for users to turn to, named Xiaoice, or Little Bing. What could be the apex of the largest Turing Test in history? 25% of Xiaoice’s 40 million followers have already told Xiaoice “I love you”. Soon Chatbots will open a symbolic door that cannot be closed again, surpassing healthcare professionals in their ability to correctly diagnose and treat their patients, simultaneously treating millions, even billions of patients at once, all the while learning, improving and waiting…
By unifying experiences across devices, platforms and OS through an agnostic interface, healthcare chatbots will soon become fundamental to our health and wellbeing. By driving adoption through the personification of chatbots – giving them endearing names, amicable personalities and aspirational avatars – we normalise their presence. But we must also be conscious of our own frailty. A dependency on or affinity to chatbots will be a real problem, and we should first agree on a new sort of 12 commandants: universal guidelines to govern our chatbot interactions.