Warning about impending ’emergency’ is manipulated by AI

AAP FACTCHECK – A US news presenter has warned that a national emergency is imminent, social media users claim.

This is false. The audio of the genuine clip has been manipulated, likely with the help of artificial intelligence technology.

In the viral video, CBS news anchor John Dickerson supposedly tells viewers a nationwide emergency alert will sound on their phones, radios and TVs in the next 24 hours, and it will not be a drill because “the government is preparing for something really, really big”.

“Political tensions are rising and global uncertainty is growing,” the broadcaster apparently says.

“This test is no coincidence, it’s a clear sign that they are gearing up for a major event. They want to ensure everything is in place for when a real emergency strikes. So stay alert, be prepared and pay close attention to the signals.

“Something big is on the horizon and you don’t want to be caught off guard. Make sure you and your family are ready, because when the real alert comes, you’ll want to be ahead of it.”

Text alongside the video in the post says “It’s very interesting that the Russians are testing their Emergency Activation System on Oct 2”.

But the clip has been manipulated. The original footage can be seen on the CBS news website, in which Mr Dickerson’s voice is notably different from the manipulated version.

The genuine report is about the first test by the US Federal Emergency Management Agency (FEMA) of a new emergency alert system during the Trump presidency, in October 2018.

In it, Dickerson actually said: “For the first time ever tomorrow, Americans nationwide will receive an alert on their cell phones from President Trump, but it won’t be a personal message”.

“It’s the first test of a national presidential alert system that will let any president issue a warning about a crisis. That could include a missile launched by another country at the US. Or a tsunami.”

AAP FactCheck referred the post to AI expert Scientia Professor Toby Walsh of UNSW Sydney, former editor-in-chief of the Journal of Artificial Intelligence Research, and the author of Faking It: Artificial Intelligence in a Human World and Machines Behaving Badly: The Morality of AI.

When asked about the motivations behind such clips, Prof Walsh said it was often just for fun, but that some people had more sinister intentions.

“It used to be that you needed some expertise, expensive software and significant [computing power] to be able to make such content,” Prof Walsh told AAP FactCheck.

“Now it is available to all, often for free and it requires minimal expertise.

“There is evidence that bad actors, often outside our national borders, are using such deepfakes to sow distrust and pursue damaging agendas”.

He said this was testing the integrity of democratic elections.

“We are now in a world where you have to ask if any image or video is real. When we no longer trust anything we see, it is a world that authors like [Aldous] Huxley imagined,” he added.

“The price is paid by those telling the truth as we no longer believe them.”

There are other telltale signs that the video being shared on Facebook has been manipulated using AI technology, apart from the voice.

Throughout the clip, the speech doesn’t quite line up with the movement of the lips.

There are also awkward, unnatural pauses and head movements, for example at the 11 second mark.

The speech is hurried at times and the intonation is unusually flat, such as at the 27 second mark.

All information, text and images included on the AAP Websites is for personal use only and may not be re-written, copied, re-sold or re-distributed, framed, linked, shared onto social media or otherwise used whether for compensation of any kind or not, unless you have the prior written permission of AAP. For more information, please refer to our standard terms and conditions.