Researcher: Algorithms Cannot Be Blamed for Disinformation, But They Contribute to It

Columbia University researcher shared her perspectives at event hosted by The Atlantic that featured former President Barack Obama.

Researcher: Algorithms Cannot Be Blamed for Disinformation, But They Contribute to It
Photo of Camille François, Karrie Karahalios and Casey Newton

WASHINGTON, April 14, 2022 – A researcher from Columbia University says that algorithms such as the ones Facebook uses cannot be blamed for causing mass disinformation, but that they must still be discussed as contributors to the phenomenon.

Camille François discussed the matter during a conference on disinformation and how it affects democracy hosted by The Atlantic magazine and the University of Chicago Institute of Politics featuring a discussion on combating disinformation with former President Barack Obama.

Through the coronavirus pandemic, disinformation surged online related to topics such as vaccines and the origins of the virus, and as war wages on in Ukraine the intentional spread of false information online has proven a chief tactic of Russia during its invasion campaign.

François spoke on a panel with Karrie Karahalios, a computer science professor at the University of Illinois Urbana-Champaign, that was focused on the power of algorithms.

During the discussion, Karahalios commented on proposed legislation in the House of Representatives to remove Section 230 protections for online content which is promoted algorithmically – thus subjecting them to legal liability. She stated that content regulation truly must be done on a case-by-case basis rather than applying a one-size-fits-all approach to regulation in all settings.

She expressed concern over some negative effects of algorithmic technology than others, highlighting as problematic the use of algorithmic technology to detect employment fraud in Michigan which led to false accusations of fraud made against several individuals.

Karahalios said results produced by algorithms such as these must be able to be contested due to their flawed nature.

Similarly, François stated that when algorithms are used for very serious practices such as criminal sentencing there must be transparency about how they are used.

To find potential solutions to some of the issues algorithms create, Karahalios suggests that data related to algorithms such as on how Facebook promotes certain advertisements be made available to a wide variety of researchers.

Earlier at the conference, Barack Obama said that he “underestimated the degree to which democracies” are vulnerable to misinformation and disinformation.

The former president said that the U.S. must mitigate the influence of dangerous online misinformation through a mix of regulation and industry standards.