Pandemic Proves to Test AI and Machine Learning Systems

Machine learning (ML) and artificial intelligence (AI) can be helpful tools in the world of data analytics and cybersecurity, helping to predict human behavior and identify potential threats. But what happens when an event causes humans to act in an unexpected way?

The COVID-19 pandemic has changed the world in many ways, down to impacting the effectiveness of ML and AI, reports Will Douglas Heaven for MIT Technology Review.

The pandemic is “Causing hiccups for the algorithms that run behind the scenes in inventory management, fraud detection, marketing, and more,” says Heaven. “Machine-learning models trained on normal human behavior are now finding that normal has changed, and some are no longer working as they should.”

Nozzle, a London-based consultancy specializing in algorithmic advertising for Amazon sellers, has been at the forefront of exploring the impacts of COVID on algorithms. Nozzle research showed that within one week at the end of February, the “top 10” lists at Amazon were full of COVID-related items across multiple countries.

“It’s an incredible transition in the space of five days,” says Rael Cline, Nozzle’s CEO.

Heaven says that the severity of the impact varies, depending largely on the purpose of the AI/ML-based program. Automated inventory systems may assume there is an error with the number of orders when they are suddenly much higher than usual. Streaming networks that have a sudden increase in viewers has impacted the accuracy of recommendations.

While machine learning is designed to be responsive, issues may arise when a new data set is too different than the one used for training, says Heaven.

“Many of these problems with models arise because more businesses are buying machine-learning systems but lack the in-house know-how needed to maintain them.”

Experts in machine learning, artificial intelligence, and data science are needed to monitor for potential breaks in the system.

Cline emphasizes the constantly changing situation makes optimization without human intervention even more difficult, and may prove to some who thought ML and AI could function wholly independently, that will not always be the case.

“You need a data science team who can connect what’s going on in the world to what’s going on the algorithms,” says Cline. “An algorithm would never pick some of this stuff up.”

Students at Capitol Tech studying analytics and data science are well positioned to help improve AI and ML systems by studying text mining and natural language processing, advanced machine learning, data visualization, and computer vision.

Want to learn about analytics? Capitol offers bachelor’s, master’s and doctorate degrees in analytics and data science. Many courses are available both on campus and online. To learn more about Capitol’s degree programs, contact admissions@captechu.edu.