Inherent in every conversation about the future impact of deep fakes is the notion that falsified visual narratives like computer-modified or generated video are somehow more powerful than false textual rumors. In this telling, a falsified video showing a politician stumbling about drunk on stage will somehow be far more damaging to their political career than a textual rumor claiming they are a habitual drunk and quoting multiple anonymous sources as having seen the drunk politician berating or even physically attacking others. In an online world where an anonymous satirical Website can launch a global story that the Pope endorsed Donald Trump, it seems we have little need of AI-generated fakery to spread digital falsehoods.
The great danger of a future world filled with deep fakes is that we would no longer be able to trust what is real and what is false in a world in which “seeing is believing.” According to this line of reasoning, deep fakes could sway elections, force political and business leaders from office, destroy businesses and livelihoods, undermine democracy and even start wars.
Yet for all this hype and hyperbole about the future potential dangers of deep fakes, it is ironic that many of the most viral digital falsehoods to date have been textual.
From the Pope endorsing Trump to Pizzagate, none of these falsehoods involved carefully Photoshopped imagery or deep faked videos. They were simple textual rumors spread by satirical Websites and ordinary social media accounts.
That text-only rumors could carry so much weight and not only influence national debate but actually spur physical violence in the case of Pizzagate reminds us that we have no need of doctored visuals to fall victim to wild conspiracy theories, deliberate misrepresentations and utter fabrications.
If a random Webpage on the Internet can make a satirical joke go so viral that it becomes common accepted truth and a tweet can cause a chain of reactions that leads to someone firing a weapon in a family restaurant, will the rise of visual falsehoods really make any difference?
In short, we don’t need digitally altered imagery or manipulated video or machine-crafted video clips to force rumor into fact. Text is more than sufficient.
The Web has stripped away the concepts of provenance and distrusting unverified content. Whereas children were once taught not to believe what they see on the Web and to only trust content from authoritative sources and even then to verify everything they read before sharing or taking action upon it, today in our rush to be first, we accept the Web at face value. We no longer question what we read or even bother to check a second source to see if there is widespread agreement or whether there are contested details.
The coming dangers of deep fakes arise not from technological advances but rather from the loss of societal information literacy. If the public was taught from the earliest age not to trust anything they see from an unknown source and to extensively verify and validate and triangulate every piece of information before sharing or considering it, we would have little fear of the coming era of deep fakes.
In the end, combating the dangers of deep fakes is not a technological issue. It is a societal one.
