The Effects of COVID-19 on Conversational AI
Business + ManagementOpinionConversational AIposted by April Miller October 21, 2021 April Miller
When the COVID-19 pandemic swept over the globe in early 2020, millions of people found themselves struggling with loneliness like never before. Mental health was drawn to the forefront of collective concern. Lockdown requirements were necessary to keep everyone safe, but they presented challenges for mental health specialists and people seeking treatment. Demand for therapists was surging while doctors were in the process of moving to telehealth services.
In the midst of these challenges, an innovative solution to mental health treatment surfaced: conversational AI. This exciting new application of artificial intelligence technology may just be a glimpse into the future of therapy.
Conversational AI and Chatbots
When people hear the phrase “conversational AI,” the automated chatbots on customer service sites may spring to mind. However, there are some key differences between true conversational AI and chatbots.
Artificial intelligence generally functions around a machine learning algorithm that develops as it gets more user input. Chatbots, on the other hand, are confined to significantly less sophisticated code. They have no ability to adapt to or learn from user input and can only generate pre-scripted responses based on keywords and standard questions.
AI, however, is capable of understanding original user input and generating a customized, original response that mimics actual human speech.
Chatbots are like trains, confined to their tracks, while conversational AI is more like all-terrain vehicles, able to go down whatever paths they need to. This level of sophistication is challenging to accomplish since computers aren’t naturally good at conversation the way humans are. Great strides have been made in recent years, with popular AI like Siri and Alexa bringing conversational AI to the general public.
COVID-19 and Mental Health
One thing is for certain: COVID-19 changed normal life forever. Some of the necessary responses to lockdown have become everyday expectations. For example, while most Americans were forced to work from home during the height of the pandemic, 48% plan to continue working from home even after it is safe to go back to the office.
Numbers like this suggest two things: concern about health and safety still linger, and many people have gotten comfortable working from home.
While there are certainly benefits to working from home, and clearly many prefer it, the lack of social interaction it causes can have implications for mental health. This was an especially big problem in the height of the pandemic and afterward in the shadow of the delta variant. During the pandemic, the number of adults who reported experiencing symptoms of anxiety and depressive disorders quadrupled from 11% to 41%. That number is significantly higher in young adults (18 to 24 years old) at 56%.
It is important to remember that these figures only represent the number of people who reported their symptoms. In reality, many more people are likely coping with loneliness, anxiety, and depression than statistics suggest.
This surge in people struggling with mental health issues was made worse by the economic consequences of COVID-19. Many people were left on furlough or completely unemployed and therefore without insurance or unable to pay for therapy. Even if they were able to afford to see a doctor, getting an appointment was more difficult than usual due to the surge in patients and the limited capabilities of telehealth. The world needed an additional solution, one that was accessible in high demand.
Chatbots have been around the internet for years. They make a handy tool for customer service centers and they’re fun to play around with, just to see what kind of response they’ll generate. Chatbots lack the ability to understand and convey human emotions, though.
As discussed above, the advanced machine learning of true artificial intelligence can actually train a computer to understand and process emotions. This technology was in development well before COVID-19 hit. In fact, it was even applied to help elderly adults cope with loneliness. The pandemic inspired the widespread use of advanced chatbots and new developments in conversational AI. One company has even gone so far as to design digital personas for its AI companions. Meanwhile, students at Emory University used Amazon’s Alexa AI to create a help bot for classmates.
While conversational AI can’t prescribe medication or treatment plans for people seeking mental health medicine, it can perform the most basic function of therapy: listening. This was the core success of conversational AI during the COVID-19 pandemic. Sometimes the easiest solution to loneliness is simply having a conversation. AI gives people someone to talk to who will be available whenever they need, for as long as they need. This simple capability, listening and responding with empathy, can make a monumental difference for many people.
The Future of Conversational AI
While AI may have lingered in the futuristic background of tech prior to COVID-19, it is now at the forefront of public interest. Surveys show that demand for telehealth will continue well after the pandemic subsides, meaning more and more people will seek help online first.
As demand grows and diversifies, conversational AI platforms will become more common and more intuitive. The more data computer scientists are able to collect from user input, the better they will be able to teach AI systems about human emotions and conversation. It will likely become commonplace for AI to also feature digital avatars, as well, allowing them to show facial expressions and physical conversation cues. While the peak of the COVID-19 pandemic may be behind us, the future is only beginning for artificial intelligence.
Learn More About NLP and Conversational AI at ODSC West 2021
At our upcoming event this November 16th-18th in San Francisco, ODSC West 2021 will feature a plethora of talks, workshops, and training sessions on NLP and conversational AI. You can register now for 30% off all ticket types before the discount drops to 20% in a few weeks. Some highlighted sessions on NLP and conversational AI include:
- Transferable Representation in Natural Language Processing: Kai-Wei Chang, PhD | Director/Assistant Professor | UCLA NLP/UCLA CS
- Build a Question Answering System using DistilBERT in Python: Jayeeta Putatunda | Data Scientist | MediaMath
- Introduction to NLP and Topic Modeling: Zhenya Antić, PhD | NLP Consultant/Founder | Practical Linguistics Inc
- NLP Fundamentals: Leonardo De Marchi | Lead Instructor | ideai.io