The alarming rise of deepfakes
“We live in a fantasy world, a world of illusion. The great task in life is to find reality.”
— Iris Murdoch (1919-1999), IrishBritish novelist
Thanks to the audio and video manipulations made possible by artificial intelligence (AI) and other emerging technologies, seeing should no longer be a sufficient prerequisite for believing.
Deepfakes are springing up from the internet’s seedy underbelly like poisonous toadstools, threatening to further undermine democracy.
On Feb. 22, the Toronto Star’s Manuela Vega reported on the fake “Joe Rogan Experience” interview between the controversial American podcast host and Prime Minister Justin Trudeau that surfaced on YouTube last month.
There is no video accompanying the manufactured conversation, just photos of Rogan and Trudeau posted side by side. It doesn’t take many seconds of audio before you doubt its veracity because the content is clearly preposterous, even if the voices do sound close to the real thing.
At one point during the exchange, the faux PM answers the faux Rogan’s question about his handling of the “trucker protest” in Ottawa by saying, “If I had my way, I would have, uh, probably nuked the capital, uh, to be honest.”
(Unsurprisingly, it may take a little longer for someone listening to the fake Rogan/ Trump interview to figure out that the outrageous conversation is contrived).
SATIRE VERSUS MALEVOLENCE
We’ve often seen the manipulation of famous people’s images for satirical purposes — in editorial cartoons, for example — and satire can be an extremely effective tool for skewering hypocrisy and deflating bombast.
But the digitally manipulated deepfakes we’re seeing today can be much more sophisticated and insidious.
And some, far from being intended as thoughtful satire, have a much more malevolent intent.
Can you imagine the impact if someone seeded an election campaign with political “advertisements” that were actually deepfakes, showing leading candidates espousing repugnant views contrary to their actual positions, or purporting to be real video showing politicians in compromising
situations?
The fallout would be immediate and catastrophic. Reputations and campaigns would be shredded and great gains made by unscrupulous opponents or their supporters.
DETECTING AND DEBUNKING
It’s a possibility that is being widely discussed in Canada by policymakers, cybersecurity experts, academics and lawyers.
In a perspective piece written for the Canadian Global Affairs Institute (CGAI) in February 2022, Abby MacDonald writes, “A realistic deepfake carefully created and released at a certain time could sway the results of a democratic election, incite violence against certain groups or exacerbate political and social divides.
“… We have not seen many deepfakes during recent elections, with altered videos of politicians and other influential people mostly being of cheap-fake quality. However, deepfakes are still improving, at a rate far faster than detection techniques.”
The question is, will detection and debunking tools be able to keep pace with the innovation churning out deepfake images that are more and more refined?
And will social media platforms be willing to put their money into the constant vigilance needed to keep deepfakes off their sites?
DO YOUR DUE DILIGENCE
As you might expect — given the wide reach of such a powerful AI tool — deepfakes are already being used as weapons of war.
Last year, a fake video quickly made the rounds on social media that appeared to show stalwart Ukranian President Volodymyr Zelensky urging the people of his country to surrender to Russia.
It was quickly denounced as a fake by sharp-eyed viewers, but how many people saw it and believed it? How much damage could it have done, both to Ukrainian military forces and civilians, if it had been more masterfully executed?
Meanwhile, deepfakes are still used most often as “a tactic to shame and discredit women,” as the CGAI’s MacDonald writes — superimposing women’s heads onto bodies in pornographic videos, for example, and maliciously sharing those images online.
For all these reasons, and because of the endless possibilities of deceitful deepfake applications, we must be careful not to accept everything we see as the truth and not to share everything we believe without doing due diligence.
The truth is still out there, but we will need to work a little harder, now, before we find it.