Paravision Deepfake Detection launches in time for 2024 global election year

The rise of deepfakes is making it harder to distinguish real information from a well-crafted AI-generated imposter, threatening the integrity of democratic processes by compromising the media that citizens depend on to make informed choices.

2024 is a critical election year around the world. This year, voters will head to the polls in at least 64 countries as well as the European Union. As many as 49 percent of people globally will participate in elections.

Paravision, a facial recognition company, has announced its deepfake detection tool to combat identity fraud and the spread of misinformation in light of the elections that make this year especially susceptible to deepfake media.

“The increasing use of deepfakes and digitally manipulated media is concerning, especially as we begin a pivotal election year,” says Paravision CPO Joey Pritikin. “In an era where trust is more vital and elusive than ever, Paravision is proud to offer a potent tool in the fight against the deceptive capabilities of advanced deepfake technologies.”

Paravision Deepfake Detection uses AI-based analysis to determine how likely it is that content was digitally manipulated with deepfake technology, outputting a score to inform automated or manual fraud analysis.

The detection software leverages “the ability to create truly diverse, properly-consented datasets, a passion for building exceptional accuracy benchmarks, and an understanding of how to efficiently deploy modern, cloud-ready software to our partners’ production environments,” says company CTO Charlie Rice.

The product was developed with support from an unnamed country that is a member of the Five Eyes alliance. It leverages a demographically diverse proprietary dataset of over 1 million images with consent.

Paravision also recently announced the release of its liveness detection, which passed an iBeta Level 2 for complying with the ISO/IEC 30107 presentation attack detection standard. It can identify masks, displays, or printouts from a single image taken from a smartphone camera.

Article Topics

biometrics  |  deepfake detection  |  deepfakes  |  facial recognition  |  fraud prevention  |  Paravision