Seeing is believing, many used to believe.
However, as we’re becoming increasingly aware in this age of artificial intelligence, reality is becoming harder to distinguish from the crude wave of fake content being generated and flooding our feeds.
Artificial intelligence (AI) has its place and, used with care, caution and proper controls, can be a useful tool. But even the world’s leading expert in digital forensics says he’s stopped trusting his own eyes.
Hany Farid, who has made a career of identifying deepfakes, told the New York Times this month that most people could no longer distinguish real photographs, voices or video clips from AI fabrications.
“Within a year or two, our whole visual system will be utterly useless.”
That’s an alarming prediction that should concern us all.
Closer to home, Netsafe has issued a warning. As reported by the NZ Herald this week, chief online safety officer Sean Lyons said that, if we as a society “start to lose our faith in what it is that we see in front of us, then that definitely has knock-on effects for all of us”.
That trust is under attack as AI-generated “news” pages and deepfake content swamp our feeds.
Lyons says he has seen an increase in AI being used to create fake news stories, which has the potential to be damaging for individuals and society as a whole.
Many of these posts appear to be legitimate and look as if they are from real news agencies, he says.
The NZ Herald article noted that many of the posts are on pages with more than 100,000 followers. They appear to take existing stories, many featuring death and misfortune, run them through AI to rewrite them, then publish them with fake images.
It’s a particularly insidious practice when you consider the families at the heart of some of these stories.
New Zealand Warriors coach Andrew Webster called this month for stronger action against AI-generated misinformation on social media after fake Facebook posts falsely claimed his wife, Emma Webster, had died.
In April, a social media news creator shared a fake image of body bags being loaded into an ambulance at the scene of an alleged triple homicide in Hastings.
Then this week, a grieving mother whose son died in a devastating incident spoke to the Heraldof the stress she endured after parts of her story were garbled by AI.
Tyler Porter’s body was found at Whangaparāoa’s Arkles Bay in July last year, 11 days after he disappeared.
His mother, Shara-Lee Porter, shared her story with the media, detailing her grief, her petition for a “reflection bench” in her son’s honour, and her concerns after a tree decorated for him was repeatedly ransacked by thieves.
Earlier this month, multiple posts about her story were published by a page using AI-generated images and inaccurate details.
The tragedy was also used in a poll-style post asking, “Should a grieving mother be allowed a memorial bench for her son?” The post featured an image of a different woman sitting on a beach.
Porter says, quite rightly, that stories like hers deserve to be told with sensitivity, care and accuracy, “not turned into generic AI content by people trying to make a money grab off someone else’s story”.
She points out that the AI stories differ from the one she shared with journalists, and people who read the posts formed a negative view of her based on incorrect and incomplete information.
Social media giant Meta, owner of Facebook, says it’s taking action and that the industry is working on ways to improve the situation.
We’ll believe it when we see it.
Meta says keeping its platforms safe is “extremely important”, and it is labelling content as “AI Info” when it is detected or disclosed.
But as we’ve seen, AI is becoming harder and harder to identify.
Tech firms created this problem. The industry needs to work harder to fix it.
While we wait for a solution to this scourge, it’s incumbent on all of us to tread with caution as we navigate the fake content minefield.