“Deepfakes” in Our Midst
Over the last week, you may have missed a miraculous event. Mona Lisa spoke.
The Mona Lisa portrait was the subject of the latest in a series of examples of “Deepfakes” — media enhanced or enlivened by artificial intelligence that present plausible versions of reality through images, audio and video. The experiment — to bring to life famous artwork and portraits of celebrities from a single source image- was described in a paper from a Samsung artificial intelligence lab in Russia. The YouTube video quickly caught fire and has been copied and shared innumerable times through social media. The question, of course, is not about the capability so much as the application. As “fake news” fears propagate, this type of technology only fans the flames of distrust. The aphorism — you have to see it to believe it — takes on a whole new meaning in a world in which artificial intelligence is making the impossible possible.
Apply this to the political process in the U.S. today and you have the makings of a real tinder box. Truth and trust are centerpieces of our democracy, not to mention our belief in the merits of a free press and journalistic integrity. What we believe to be true and what is actual truth seem more and more at odds. Whether it’s the Mona Lisa or Nancy Pelosi ( also the subject of a deepfake exercise this past weekend), we are quickly learning that perception is not always reality. So who can you trust?
As social media, digital media and, well, all un-licensed media attempt to disseminate news of all flavors and varieties, the veracity of the information being shared is bound to come under scrutiny. Citizen journalism has the benefit of being inexpensive, “authentic,” and even corroborating/challenging to a particular point-of-view coming from governments, corporations and institutions. It’s a check and balance of a certain class — but it’s not a gold standard and it doesn’t hold itself to any set of ethics. There’s no accountability for citizen journalism. So as we plunge headfirst into the age of Artificial Intelligence, it’s wise to question any form of media being presented by unknown authors with unknown agendas.
Broadcasters, in contrast, are licensed by the government, and that may be more important today than ever before. Local broadcasters are licensed to serve the “interests of their local communities.” Any content that violates or offends the community can put a local broadcaster’s license in jeopardy. In other words, what plays in Peoria doesn’t have to play in New York. The mores of each local community therefore drive the type and manor of content that broadcasters provide — including strong repercussions for willfully misleading or misrepresenting the facts. There are also laws that protect the opportunity for multiple voices or points-of-view to be heard in each market. This regulation sets broadcasters apart in the coming age of Artificial Intelligence. Cable operators, digital home pages, recommendation engines and social media platforms have no such restrictions or repercussions for being bad actors in our communities. Just this weekend, it took Facebook 32 hours to even address misleading video disseminated on their platform about Nancy Pelosi, the United States Speaker of the House of Representatives. Although now “downgraded” in its algorithms, the video continues to be shared and includes only a text warning about entities that are actively investigating the video clip’s origins.
New York Times, May 26, 2019
In response to the viral video, Facebook’s product policy and counterterrorism executive, Monika Bickert, stated, “We think it’s important for people to make their own informed choice for what to believe,” she said in an interview with CNN’s Anderson Cooper. “Our job is to make sure we are getting them accurate information.” Here we have “accuracy” and “believability” at odds in essentially the same explanatory response.
So while broadcasters are bound by their legal obligation to serve their communities, digital entities like Facebook are in actuality shielded from responsibility thanks to Section 230 of the Communications Decency Act of 1996. Section 230 gives immunity from liability for providers and users of an interactive computer service who publish information that is provided by others. Originally intended to spur innovation and encourage start-ups, the law is now being used by some companies to evade responsible dissemination of information.
Under a political lens it may be difficult for some to see the forest for the trees here, but substitute the House Speaker for your teenage daughter or your aging Granny and the issue gets personal pretty quickly.
In the interest of fairness and balanced representation, however, you can read an alternate opinion on Facebook’s decision from MIT in this Ethical Tech article published earlier this week. The argument focuses on free speech and the dangers of deciding which voices can and cannot be silenced. The issue here, however, is not about giving a forum to dissent so much as it is to propagating disinformation. For decades, local broadcasters have been pretty good gatekeepers, protecting everyday citizens from slander and biased representations. Real journalism is concerned with finding the “truth” and presenting facts that are corroborated by multiple sources in an objective way. This does not mean that broadcast news organizations always get it right. Mistakes can and do happen. But more often than not, broadcast news leans toward factual, objective storytelling rather than sensationalized click-bait headlines or salacious on-air debates. This is why local broadcast turns up on top of every poll among among U.S. citizens when it comes to “trust.” Local TV and radio stations — depending on the poll or pollster — consistently are tops with viewers and listeners. This is no small anomaly either. Poll after poll, survey after survey, rank local broadcasters as most trustworthy and least-biased.
It’s no secret that digital media, tech sectors and artificial intelligence are all galloping toward an automated, realist- ic existence — but the key is in the “ ic.” Privacy is already under siege. When we can no longer trust what we hear and what we see in digital life — tangible, local and experiential interactions with media and brands will be the ultimate deciding factors of the “real” — with no prefixes or suffixes (sur-, -ic, -ism, etc) to dis-intermediate our impressions and decision-making. Brands wishing to establish themselves as authentic, must begin to engage locally now to get a foothold, and alignment with local broadcasters will be key to validating their credibility.
Accept no forgeries, no simulated engagements nor “deepfake” chicanery. Connect authentically with local media partners who have made it their business to accurately represent and engage their communities. Come back home to local broadcast.
Originally published at https://blog.katzmedia.com on May 28, 2019.