UC Berkeley students launch browser extension to combat fake photos

surfsafe_surfsafe_courtesy
SurfSafe/Courtesy

Related Posts

Two campus seniors, Rohan Phadte and Ash Bhat, recently launched a free browser extension called SurfSafe that detects fake or doctored photos to combat fake news online.

When installed, the user hovers over the image in question, and SurfSafe instantly cross-checks it with images in its database to determine if there is a match, according to Kyle Rentschler, a SurfSafe product engineer and campus senior.

Rentschler added that as users browse, SurfSafe creates a “hash” of each image that is added to its database for future matches.

“It sends the URL of the images that you see online to our back-end server, and what we do is we essentially compute what you can think of as a like barcode or a fingerprint for an image,” Rentschler said. “That’s what a hash is, and it allows us to quickly and efficiently look it up and compare it with other images in our database.”

There are still some limitations to the extension, Rentschler said. If a user tries to verify an image SurfSafe has yet to encounter, the plug-in cannot determine whether it’s fake or real, and, instead, it notifies the user that there are no matches.

Another limitation, Rentschler added, is if a user is on a website with too many images or scrolling too fast, the plug-in will not work. In cases like this, Rentschler said users can manually send in requests to cross-check a specific image with all the barcodes in its database. SurfSafe matches the image perfectly or with similar images and sends the user a list of where else the image was found online.

For some, however, SurfSafe is not enough to deal with misinformation on social media. According to UC Berkeley’s Graduate School of Journalism’s Assistant Dean for academics Jeremy Rue, users should understand how misinformation campaigns operate and think critically about everything they are sharing, even if they agree with it.

Nihar Dalal, a campus senior and computer science major, said relying on an algorithm as SurfSafe does can lead to issues with validity.

“(SurfSafe) runs on an algorithm, and the algorithm is not always perfect,” said Dalal. “Even if (there is) one time when it marks real news as fake, that could be like pretty detrimental, because that would just spread throughout and … just add to the problem of fake news.”

The plug-in is available only for Google Chrome, and the team is working on making SurfSafe available to Opera and Microsoft Firefox users, according to Rentschler.

Rue said that while SurfSafe is a good tool to address the issue of misinformation online, consumers must remain critical of their social media.

“There’s no one arbiter of what is trustworthy and what isn’t,” said Rue. “I think it’s just being transparent so people can make those decisions for themselves.”

Contact Mani Sandhu at [email protected] and follow her on Twitter at @ManiSandhu24.