American Univesity of Beirut

Can AI be used for good?

​​​​​​​​​February 28, 2021​​

​Dr. Fatima Abu Salem will be speaking at the Women in Data Science ​(WiDS) Worldwide Conference​ on March 8, 2021


How do you use AI in your work?

I am most interested in applications for the public good, rather than mainstream applications of AI such as in robotics or other fully automated services. My work employs machine learning, which is a sub-set of AI. We use a suite of modeling techniques that learns from data to answer questions we have about crises affecting people and motivated by their behavior.

What motivated you to become involved in this field?

I grew up during the war in Lebanon, which had an impact on our family. I was also of refugee status and became acquainted with discrimination. When I got involved with mathematical skills and data science, I thought maybe I could use my knowledge to better understand the circumstances that impact people in conflict.

Give an example of the kind of projects you work on 

During the Syrian war, it seemed natural to ask how to identify damaging fake news concerning the war. We also looked at demand from Syrian refugees on primary health care centers in Lebanon. When do health visits peak, and do they correlate with peaks in the war, during which mobility is perturbed? Is there an association between birth defects and air pollution? How did the political stance of newspapers in Lebanon change over time with respect to the Israeli-Palestinian conflict? How can irrigation metrics be predicted more accurately in order to improve on water usage in farming? We have hunches about how things work, and we fit models using available data to allow computers to predict such important questions. The outcomes from these projects can help policymakers and others prepare for future trends and events.

What are the top challenges for women here in data science?

In Lebanon, political instability is the most difficult. In addition to our work, we have to worry about our livelihoods and our future as humans and mothers. Funding remains an issue, especially in a developing region. And gender stereotypes still play a role. We still find ourselves having to try harder to prove that we’re good enough. When men support their female colleagues, it makes a tremendous difference, and we’ve made progress in this regard. Otherwise, it can be very lonely.

Are there specific challenges to machine learning in Lebanon? 

A machine learning-based, scientific and industrial culture cannot thrive without data, and our country has not sufficiently embarked on the data revolution nor has it been as transparent as it could have been in this regard. The number of disasters we face hinders rigorous data gathering. We do have sizable data generated by NGOs and other agencies, but people still exercise monopolies over it. That’s difficult when we’re trying to do a service, but we use work-arounds to maximize learning from what we do have. These experiences compels one to hone their analytical skills further and further. 

We always need more cloud storage and dedicated software, and financing that is hard, but we have a rich research culture in Lebanon that I hope we’ll be able to maintain. We’re constantly engaged in the most recent technological and scientific developments. I think the Lebanese milieu has been impressive in that sense. 

From a human perspective, there is zeal to embrace technology here, but not the same zeal to rigorously develop the technology. The volatility of the region can breed a transient mode of thinking, which is why many projects are short-lived. 

What are you most excited about right now in your wor​​k?

We have a grant from Google’s AI Impact Challenge for the Social Good that we are applying to irrigation issues in agriculture, and we’ve generated some very promising results. This has given me a big boost and propelled me into unchartered territories in the kind of predictive modeling that should at the same time meet standards for trust by the end user. 

The field of machine learning is pervasive in a beautiful fashion, and offers hope that many daunting problems can begin to be solved by it. The variety of data nowadays and the solid mathematical foundations behind machine learning can impact unlimited applications. I believe there is room for many people of all backgrounds to contribute and participate for years to come. 

Contact Us

For various questions, please try contacting us via social media first!
read more

Privacy Statement

We take data privacy seriously and adhere to all applicable data privacy laws and regulations.
read more

Copyright and Disclaimer

Written permission is needed to copy or disseminate all or part of the materials on the AUB website.
read more

Title IX, Non-Discrimination, and Anti-Discriminatory Harassment

AUB is committed to providing a safe, respectful, and inclusive environment to all members of its community.
read more