‘Insidious’: Why this AI trend needs to stop
There is a very dark trend making waves online that experts say spells disaster for our future.
Red Carpet
Don't miss out on the headlines from Red Carpet. Followed categories will be added to My News.
COMMENT
Well, well, well. Look what we have here: A new Pope (Leo XIV), new golly-we’re-on-the-brink-war fears (between nuclear powers Pakistan and India) and a new/old Prime Minister whose most pressing job today is sampling bonbonerie and planning his bucks to Coffs.
Then in New York, we got a new appearance by Katy Perry on the Met Gala’s shocking blue (blue I tell you) ‘red’ carpet.
This year’s extravaganza of laboured sensationalism was as usual the sort of contrived outré stunting we have to expect, an exhausting grasping of celebrities scambling to make best dressed lists, but there in the midst appeared to be Perry, wearing a futuristic, sculptural suity thing. It was stunning.
It seemed like a smart play. Fresh off of her being pasted and pasted some more for using her precious minutes of zero gravity during her Blue Origin space flight to plug her world tour, here was 40-year-old looking sleek, contemporary and cool.
Nothing like a bit of red carpet image laundering, no?
Except that none of it was real.
On Monday evening, New York time, as images streamed out from scene on the Met’s front steps, this shot of ‘Perry’ did the rounds to rapturous applause, before someone clocked the truth - the California Gurls singer had not beetled to New York to pose like her career depended on it but was in Texas.
This supposed Met Gala ‘shot’ of her was the magic fever dream of all the ones and zeros of some code in Silicon Valley - it was nothing but AI.
Then, before anyone could even begin to type out the word ‘harbinger’ and do some fretting, only two days later came a horrifying example of the darkness of AI-generated ‘news’.
On Wednesday, a completely fake video appeared online purporting to show one of French First Lady Brigitte Macron’s former pupils ‘revealing’ that she had allegedly sexually assaulted him at the age of 12.
The AI generated video had already been viewed nearly ten million times. The clip supposedly shows Lionel Torres, one of Mrs Macron’s literature students from the 1980s, detailing her kissing him and removing his shirt when he was a child - except of course it’s not real.
The real Torres denied the horrible allegations to France 2 television, saying “that’s not me and that’s not what happened to me”.
Menahile, French security officials have accused Russia of being behind the creepy video fake, alleging it was created by their Storm-1516 cyber warfare unit. (The Kremlin, according to the Times, has “hundreds of influencers” on their payroll.)
What the Katy Perry and Brigitte Macron situations expose is nothing short of terrifying - an age in which it is increasingly difficult to discern between real and completely bogus but realistic seeming news.
Seeing, since the invention of the daguerreotype, has always meant believing but that’s an axiom we can well and truly bin now.
Like that poor bloke stuck out on the prow of the Titanic getting frostbite in his extremities, it’s time for someone to desperately start ringing the alarm bell.
Welcome to the uncanny valley, because we all live here now. AI ‘news’, either generated for weaponised ends or just for sh*t-stirring funsies, is here.
This week’s crescendo of trickery comes after a series of ‘photos’ have hoodwinked the internet and the press with increasing frequency in recent years.
In March 2023, tens of thousands of Twitter accounts posted a shot of Pope Francis wearing a Balenciaga puffer jacket, along with on Reddit and Facebook with many fooled.
In the months that followed, a ‘mugshot’ of Donald Trump did the rounds (though a real one was taken and came out later) and thousands of social media accounts posted about an ‘explosion’ outside the Pentagon and caused a brief dip in the stock market.
Last year French farmers revved their tractors and packed a baguette lunch to protest Elysée agriculture policy changes and ‘staged’ a massive hay protest in front of the Eiffel Tower.
Earlier this year, after a series of American air disasters, an image did the rounds purporting to show a Delta airlines flight having crashed.
When the devastating fires broke out in Los Angeles in January year, within 24 hours of the disaster, disturbingly ‘real’ images like of the Hollywood sign on fire started circulating.
So convincing were they that Hollywood Sign Trust chair Jeff Zarrinnam told NPR that he was “inundated” with ‘too many emails to count’ from “news agencies around the world” wanting to know if the LA icon was okay.
He has said that the ‘fire’ shots “look so real that I couldn’t tell if it was real or not. You know, if I didn’t see the Hollywood Sign myself … I would have probably believed it.”
What seems clear after the Katy Perry and Brigitte Macron instances this week is that this sort of visual bamboozling is only going to happen more.
This is just a taste of what is to come and thine eyes are only going to deceive us and deceive us some more.
While there are a growing number of tools being used by the media and news agencies to detect AI, there is still a lag time between something going viral and it being outed and we are all at risk of falling victim.
Also, as the machines learn and learn fast, the sophistication and level of detail they will be able to render in images and videos is only going to get richer and even more cunningly illusory.
The timing of these technological leaps and bounds could not be worse - when trust in the established, professional media is under attack.
Mulder and Scully taught us that the truth is out there. If only we could actually spot it.
Daniela Elser is a writer, editor and commentator with more than 15 years’ experience working with a number of Australia’s leading media titles.
More Coverage
Originally published as ‘Insidious’: Why this AI trend needs to stop