How AI can facilitate preliminary medical work through communication
Artificial intelligence (AI) is already deeply anchored in the health sector. Like many other industries, the health industry has to deal with huge amounts of data, which can no longer be managed manually. In addition, the learning capability of software offers certain advantages and strengths, which a single person cannot offer on their own. For example, machines can discover and analyse abstract patterns in X-ray images and body fluids such as blood or urine much faster than doctors. In addition, through each “activity” the software gathers more data and improves itself to achieve even more accurate results in the next task.
Chatbots in the health industry
Market researchers expect the market volume for AI applications in all sectors to rise from 1.25 billion to 36.8 billion US dollars by 2025. The use of AI includes chatbots. The main use cases are to increase customer satisfaction and to carry out simple but time-consuming activities. One example for a specialised HealthTech chatbot is Melody, an app from the Chinese search engine Baidu for patients who want to get quick medical answers or even diagnoses. However, chatbots can also help on a smaller scale, for example in the management of doctor’s appointments.
Certain patients will always prefer human contact over communication with a chatbot, but the big advantage of chatbots is that they can be reached at any time. And not only that, chatbots can also provide crucial help in areas such as mental health. Currently 350 million people worldwide suffer from depression, one in every tenth person. According to the World Health Organization (WHO), suicide due to depression or affective disorders will be the second most common cause of death by 2020. Currently, they rank fourth. However, therapy places and treatment facilities are rare and waiting times for a first meeting can take up to six months.
Woebot to the rescue
Psychologists and AI-Experts from Stanford University developed a new mental health chatbot to address exactly this problem and provide more immediate help to those in need. Woebot is accessible to everyone via the Facebook Messenger and offers first communicative help with anxiety or depression. A first survey with 70 students showed, that the ones who worked with the chatbot showed mental health improvement after the experiment. And what is so special about Woebot is, that its “brain”, or rather its AI, differs from other health chatbots in the sense that it doesn’t need a database or similar sources to feed it. It works purely on the user’s input to function and learn.
As a big chatbot fan, my colleague Paulina Wojcik has tested Woebot for one month and provides a more detailed insight into how it works:
Woebot – A self-experiment
When I first heard about Woebot, I immediately became curious. All chatbots I knew up to this point were used only for sales or news. What fascinated me was the idea that an AI could evaluate the state of my mental health and give me appropriate tips. Immediately I asked myself: How can a robot do this? A human is certainly more capable, isn’t it? How is this supposed to work at all? In short, I decided to try Woebot for myself and was positively surprised.
From the very first “Hey Paulina!“ Woebot comes across as extremely sympathetic and makes the communication with it very easy. Right from the start, the rules are clear: it is an AI, some privacy policies apply and the small digital helper makes it clear that it is not a substitute for real help. Nevertheless, the chatbot is especially suitable for timid people who may have problems with help-hotlines. It helps to overcome the first barrier and gain comfort in talking about mental problems.
Once the basic rules are clarified, Woebot explains to me how it works and what it plans to do. It’s a long process, in which it wants to get to know me first, just as a real psychologist would. Daily check-ins allow the bot to create a pattern about my mental state (provided, of course, I answer the questions honestly).
The exciting thing is, Woebot not only asks how I am and what I am doing on a daily basis, but also engages me in psychological word games, which in turn provide additional data about my condition to help me. I learn more about myself, my behaviour and the psychological explanations for how I’m feeling. Additionally, I get tips on how to get rid of bad habits or overcome mental stress and anxiety.
In addition to check-ins and games, the small robot also actively educates me about mental health. I regularly receive content on the subject of self-help, self-confidence and stress-relief. Astonishingly, Woebot’s content often responds to my moods. One day, when I stated that I was stressed and unmotivated, I received an article on the most relaxing song in the world, with an appropriate link. After giving the content a negative review, I was immediately asked what I prefer to do to relax. The next time I mentioned being stressed and unmotivated, Woebot didn’t send me a song, but a reading suggestion. He had “remembered” that I prefer reading to relax.
After a month, however, the surprise came: Woebot wants money! A subscription to continue accessing the entire content spectrum. Only the basic features, such as the daily check-ins, remain available for free. However, a pay model makes sense at this stage, because chatbots like Woebot require a lot of work during setup. The money is needed for the optimisation and finalisation of the little helper in order to perfect him for his task as a digital mental health supporter.
At the end of the month, I had not received a proper diagnosis from Woebot – at least not in the free version, which I used. But perhaps that doesn’t matter. The chatbot helped me to understand myself and my moods better and to deal with stress, lack of motivation and mental fatigue – all through communication, gamification and education. Chatbots will never be able to replace human help in the health sector completely, but they can help overcome barriers and educate patients about their conditions. We are keen to see where the journey of chatbots in the health sector will lead.