4 minute read

AI in Healthcare

Artificial Intelligence has the potential to make strides in healthcare, but scientists must balance the optimal outcomes with the need for data privacy.

BY HAILEY MINTON

X-ray diagnostics, contact images of tracing, computer vision, molecular chest X-rays machine learning are a few Artificial and generated Intelligence tools highlighted by algorithms to Lauren Pfeifer, a Data Scientist efficiently detect and Venture Capital Investor at if someone has Maschmeyer Group Ventures, based COVID-19. in San Francisco. The difference The X-ray shows between Artificial Intelligence and nodules that look Artificial General Intelligence is like glass shards the difference between programing which indicates a bot to perform a specific job that the patient versus having a machine emulate needs further a human in the way it performs a testing. These task. Lauren says progress is moving algorithms can forward primarily with Artificial sift through vast to see where people were contacting Intelligence in Healthcare. There is amounts of X-rays and flag the ones others. Of course, there is a give AI a lot of potential for good that can come from an algorithm processing vast amounts of data because it can help pinpoint problems faster and get care to those who need it more effectively.

Computer Vision

Early detection of skin cancer can prove the difference between a simple mole removal or several rounds of chemotherapy.

AI is making strides in diagnostics of X-ray images. A handful of medical institutions have gathered images of chest X-rays and generated algorithms to efficiently detect if someone has COVID-19. The X-ray shows nodules that look like glass shards which indicates that the patient needs further testing. These algorithms can sift through vast amounts of X-rays and flag the ones that resemble the X-rays of patients with COVID-19. The turnaround time is the bottleneck that has been optimized.

Contact tracing is another AI tool that has been implemented in Israel to find the likelihood of COVID-19 outbreaks in certain parts of town. They use geolocation and tracking to see where people were contacting others. Of course, there is a give and take with data privacy. This leads us to ask the question, what data is being used to make our lives healthier and what is it worth?

Computer vision is another branch of AI that can be very powerful in the realm of infectious diseases. Zap Malaria is based in Israel and they use geospatial data to find bodies of water around areas experiencing malaria outbreaks. Using satellite data, they look for ideal breeding grounds for malaria-carrying mosquitoes and render the information to field workers. The field workers travel to these bodies of water to confirm, after which they dispatch doctors, mosquito nets and antimalarial drugs to that area.

To Lauren, the field of molecular machine learning is the most exciting frontier of AI in healthcare. “They mapped out the immune system to run simulations to to test drugs.” If a person has several conditions, they can give a simulated outcome to know if a drug will work. Pfeifer hopes these solutions can be applied to the demographics that need it most.

DATA PRIVACY

The strides that can take place with AI in healthcare depend upon large amounts of data to make it effective. But understandably, people might not want their data, especially sensitive health information, being used. We have HIPPA and data privacy as safeguards to protect people, but that barrier limits the data that complex algorithms can compute. Just like each of these examples, the output of these algorithms can save lives, but it is dependent upon the input. There is, however a flip side: Data Privacy.

There have been unforeseen consequences of using AI in collecting and analyzing vast amounts of data. Pfeifer referenced the Netflix movie “The Social Dilemma” which addresses the unintended effects that social media has had on mental health and the spread of misinformation. This example helps us feel the weight of the unintended consequences that can come from utilizing AI with information that is even more sensitive than the information gathered through social media usage.

Another issue in data privacy comes with user agreements. Pfeifer says, when a person downloads and uses an app for the first time, most users agree to the policy terms by checking the box and moving on without reading them. That puts the responsibility on the user to know what data the app uses. However, Pfeifer says, “People need a law to degree to understand some of these data usage policy agreements.” To the average person, it’s like trying to understand a foreign language. Pfeifer hopes to see a push to make user agreements understandable to the average person so they know how their information is being used.

Lauren Pfeifer, Data Scientist and Venture Capital Investor at Maschmeyer Group Ventures

HOW CAN AI HELP HEALTHCARE?

There is a lot of potential for good that can come from an algorithm processing vast amounts of data because it can help pinpoint problems faster and get care to those who need it more effectively.

This article is from: