Audio manipulation threatens productive dialogue

Spencer Buckner

In an age of media manipulation, the American people have been struggling to figure out what’s real. With realistic-looking photos of President Donald Trump saving cats and leading boat rescues in Houston after Hurricane Harvey (he didn’t) circulating the internet as authentic, technology’s power to warp the fabric of reality continues to reaches new heights. While these images are at best humorous and at worst misleading, they represent the tip of the iceberg of media manipulation.

Adobe’s new VoCo program promises to be the equivalent of Photoshop for audio recordings. Using only 20 minutes of recorded voice audio, it can realistically synthesize a subject saying anything a user can type into the program. While this technology was demonstrated almost flawlessly with humorous effect using a conversation between famed comedians Keegan-Michael Key and Jordan Peele, VoCo’s impact on truth-finding is no joke.

In both the public and judicial sphere, audio is a critical source of finding and establishing truth. For instance, when then-candidate Trump’s infamous “Access Hollywood” tape emerged, it was undeniable that it was Trump’s voice on the tape. In the courtroom, audio can be a critical aspect in proving a defendant’s guilt — or solidifying their innocence. The credibility of audio is recognized because technological limitations have historically rendered it nearly impossible to realistically manipulate — limitations which are now disintegrating.

Imagine if a Voco-fabricated tape emerged of what convincingly sounds like a conversation between a top Trump advisor, or even the president himself and a Russian official colluding over the 2016 election. A tape like this could not only be a political weapon against the president and perhaps result in legal consequences. But if discovered to be fake, it could also render a chronic distrust toward using audio as reliable evidence of guilt. Courtrooms in particular have a stringent set of guidelines to ensure the use of legitimate audio, meaning that innovations such as Adobe’s VoCo could disrupt proceedings in which determining the legitimacy of audio evidence becomes too difficult.

Thankfully, Adobe’s developers are considering placing an auditory watermark on VoCo-edited audio to help ensure consumers and courtrooms alike know the difference between real and altered content. While promising, safeguards like this have traditionally been easy to crack and thus highly unreliable. Regardless of its effectiveness, a simple watermark doesn’t change the fact that this software and programs like it will no doubt be used to confuse Americans and sow distrust in officials and institutions — one of the key strategies that highly partisan dealers of fake news utilize today.

As writer Maria Konnikova noted in an article for Politico, “When we are overwhelmed with false, or potentially false, statements, our brains pretty quickly become so overworked that we stop trying to sift through everything.” With new programs like VoCo making lies look more convincingly like the truth, the sifting is going to get a lot tougher — and the manipulation of the truth a lot easier.

In a society where people are disagreeing on facts and not just opinions, having a clear and informed grasp of the truth is vital in ensuring productive dialogue continues to exist. Developers of VoCo must understand the power they hold over the truth itself and thus ensure that consumers know when audio has been manipulated. Adobe’s responsibility to safeguard the truth in today’s society cannot be overstated, especially when our perception of reality hinges on it.

Buckner is a Plan II and government freshman from Austin. He is a columnist.