On April 1, 2022, the GreatFake occcurred.
It started at 3:30AM Eastern Time, when 17,985 state and local police departments began receiving a slightly elevated number of phony 911 calls. These calls took the form of local residents, panicked about a variety of fantastical things. Lights in the sky. Weird sounds and voices from their computers. Sometimes it was them reporting some terrible crime they had just seen another local commit. Other times, committing suicide on the phone call.
These were not local residents, of course. These were human controlled networks of AI bots, using tens of thousands of local personas crafted by mining social media for voice and speech patterns. They were the voices of real people, and could answer simple questions about the people they were posing as. They could easily pass any Turning test a harried 911 officer could concoct.
We simply had no defense for this. As the calls escalated, by 5:30AM the call centers were jammed to a point where the Federal Government was on the case.
Unfortunately for the Federal Government agencies, they had been hit by a similar assault. But this time it was AI posing as people’s bosses calling them and telling them not to come into work, that they were fired, or committing suicide on the phone.
It quickly became clear that every government and police official in the country had an AI version of him or herself running amuck. These AIs were sending offensive emails to the press, deep-fake nude photos of colleagues, and locking police and emergency responders out of critical systems.
On YouTube, deepfake versions of celebrities spouted conspiracy theories in crafted videos. The president was shown doing terrible things, but nobody really cared. In fact, what folks had missed all along was that the threat deepfakes posed wasn’t to celebrities, it was to them. Their everyday lives.
What the nation awoke to on that April 1 was a complete breakdown in trust of voice and video. The repercussions were profound. Thousands died. Cities rioted. Fear of bots spread like a cancer. Social media networks were shut down by the government.
And the perpetrators were only doing it to short-sell stocks in American markets, and buy cheap on the bounce back. A bunch of gangsters. In this case, Russia was pissed because they stole the plan it had saved for a rainy day.
But then something happened. We cleaned up the mess and decided to fight back. Systems for controlling your identity and validating content creation quickly took hold and flourished. There were false starts, but within five years every human on the planet who could afford an electronic device, which was most of them, could validate any video or voice communication of themselves easily and securely. And without that validation, video and voice communication was instantly rejected as compromised.
We whitelisted ourselves.
The GreatFake, like the Thanos snap, was inevitable. The only question was what form it would take. In another, darker world, perhaps the attackers went after our nuclear facilities.
We were lucky.
Image from 3D printed mask to help you evade surveillance, project here.
Story inspired somewhat from this research to put words in your mouth.
By analyzing the speech patterns in a video of one person speaking into a camera, they’re able to swap whole phrases for completely different ones. The result is pretty realistic.