UAB professor part of group awarded grant to research election ballot security

A UAB professor is part of a group researching voting ballot security.
A UAB professor is part of a group researching voting ballot security.(WBRC)
Published: Oct. 8, 2022 at 8:12 PM CDT
Email This Link
Share on Pinterest
Share on LinkedIn

BIRMINGHAM, Ala. (WBRC) - The National Science Foundation awarded a group, including a UAB professor, a grant to research election ballot security.

We are only one month away from the 2022 general election here in Alabama and with every election comes the discussion about election security.

Most elections rely on hand-marked ballots where voters fill out a bubble target beside the name of who you want to vote for, but the technology that reads those ballots are not always accurate, according to Chengcui Zhang, a UAB computer science professor.

“I think everybody wants their vote to be counted, right, and to be counted correctly,” she said.

The four-year $1.2 million grant will help Zhang, along with a professor from Rice University and Texas A&M, research how to create a more accurate system when counting ballots.

“In some cases, you might see a ballot with marks outside of the bubble target area,” explained Zhang. “For example, a tick mark next to the candidates name. As a voter, I want that to be counted, but the existing scanning system will count it as a non-vote.”

That’s where the project comes in to enhance the system. Zhang describes the two main purposes for it below.

The first: “Develop an AI assisted tool to compliment existing ballot auditing system by better recognizing the hand-marked bubble targets more accurately.”

The second: “Detect potential fraudulent cases. For example, the same voter fills out multiple ballots. Our preliminary study has shown that it is possible to detect those voting frauds by looking at the official similarity or the similarity between the marking styles of two different ballots.”

She says the tool could be used for any bubbling system like standardized testing, in addition to elections.

Zhang says the entire process is expected to take around 4-5 years because creating the new AI tool is one thing, but convincing people to use the tool is completely different.

Get news alerts in the Apple App Store and Google Play Store or subscribe to our email newsletter here.