How AI Can Help with Vaccination

This is the first blog post on AI and vaccination. Also see:

Who Gets the Vaccine First? AI Helps but Humans Decide

AI for Managing Vaccine Supply Chain

Covid Vaccine needs Blockchain Transparency; U.K. Sees the Light

The world got mobilized for an incredible COVID-19 fight. The first vaccines are out. Now the entire planet is facing three huge challenges – manufacturing, distribution and administration of the vaccine. Leaders and managers need to make decisions they never imagined before. The AI community, from startups to tech giants, is contributing to this enormous effort. Two main areas where AI can help are:
1) reducing uncertainty and
2) performing repetitive human tasks at scale.

AI Reduces Uncertainty

There is no shortage of uncertainty in the pandemic!

• Who gets vaccinated first and when the vaccine will be available to others?
• What is the efficacy of the vaccine? Will it degrade over time?
• How will vaccination affect the economic reopening – in different cities, states, countries?
• How much PPE do we need and where?
• What do people need to know about the vaccine effects?
• What should the governments do (and how, and when, and with whom), and what should be done locally?

These are all questions that I know for sure AI is already addressing. The timing of a return to “normal” (whatever that will be) is murky: As of December 2020, 61% of HR executives didn’t know when their organizations will resume business travel and 48% weren’t sure when their employees would reintroduce in-person client meetings or conference attendance.

Uncertainty has many levels – individual, organizational and societal. Each one of us participates at every level. Make no mistake, reducing uncertainty is not the same as eliminating it. This is the root of the AI criticism: one misstep – and AI is out the window for adopters. But being even a bit more certain is an advantage. For example, at the beginning of the pandemic, when we knew very little about the virus, the best data scientists out there were building models of the COVID-19 spread. First, the outlook was imprecise (although useful). With more data, data scientists were able to predict the disease spread pretty darn accurately. Now, the same people are working on predicting the effects of the vaccine(s), but again, we have very little data. With more data (soon!) we will get better forecasts. Action item: collect the data, share it, put it together.

Meanwhile, existing AI solutions are being swiftly repurposed for vaccine distribution and administration – this reduces guesswork, for instance, for site-based vaccine needs based on COVID-19 spread and waves. AI-enabled supply chain optimization, inventory management, logistics, demand prediction and facility staffing are just the most prominent examples.

AI Performs Repetitive Human Tasks at Scale

Vaccine manufacturing, distribution and administration include a myriad tasks. Some of them support business functions, such as order management, patient scheduling and triage. Some others are less obvious but crucial for success of the vaccination – earning trust of the population, reducing anxiety and quickly answering frequently asked questions.  These are typically considered human tasks, but AI can help with a lot of them. AI can work 24×7, it does not get tired and more importantly, it does not get frustrated.

AI language services, available off the shelf for rapid implementations, scale humans. They include speech to text or automatic speech recognition, text to speech, natural language understanding, natural language generation, translation, sentiment analysis (emotion AI), text analytics and chatbot virtual assistant frameworks / dialog management. For example, chatbots and AI-enabled contact centers answer questions about the vaccine, its side effects, health, wellness, timing and finance – in many languages!  Sentiment analysis helps surface general feelings about receiving a vaccine, including within specific communities or geographies. Then, as vaccination rolls out, automated sentiment analysis can help to track changes in beliefs and behaviors.

Computer vision technologies are another force multiplier. Image recognition supports medical imaging analysis, diagnostics and patient triage for conditions related to the vaccine and to the virus. Video content analytics enables all kinds of monitoring, including patients’ vital signs and the state of facilities (even counting people in the crowd). Optical character recognition (OCR) is a great scale amplifier for processing various medical forms.

How to Think of AI

The best way to think of AI is how to make people more confident and organizations more productive. Employ AI to compensate for human limitations, for example, by offloading repetitive work to AI and by giving people more clarity about the vaccine. Encourage humans to expand the possibilities for AI, for example, use AI for distribution, storage and funding needs. Find where machines do their best and where people can be at their best – and allow both sides to shine in the fight against COVID-19.

Gartner Coronavirus (COVID-19) Resource Center

Follow Svetlana on Twitter @Sve_Sic