In fast-deteriorating social media spaces, the tech giants cannot be trusted to do the right thing, only the thing that will make them more and more money. They seem to think irresponsibly unleashing unproven AI will help that quest.
Subscribe now for unlimited access.
$0/
(min cost $0)
or signup to continue reading
Impossible to miss as the search tool in some of the world's most popular social media platforms - Facebook and Instagram - this so-called artificial intelligence "Meta AI" has been immediately found wanting.
Not only is it proving useless in helping people who "get their news from Facebook" discern what is real and what is not, it is actively creating misinformation and falsehoods, an analysis by ACM has found.
When asked to summarise a story about the Bondi knife attacks Meta AI came out with the stunning revelation the attacker Joel Cauchi, who was shot dead by police at the scene, had "died by suicide in 2020". Really? As inaccuracies go that's hard to beat.
When asked to summarise a Canberra Times court report, Meta AI botched it. The subject of the case in question had been accused of a crime and granted bail. He had not yet entered a plea. Meta AI said he had been found guilty and would be sentenced at a later date.
If that mistake had been made by Nine, Seven, News or ACM a flurry of defamation concerns notices would quickly follow.
Meta, which argues it is not a publisher - and will stop paying Australian media organisations for using their content when the existing agreement expires this year - believes it can get away with disseminating such misleading rubbish scot-free. That's despite the potential to seriously harm the reputations and community standing of real people.
It's only a matter of time before somebody posts a Meta AI-generated libel on their Facebook or Instagram feed and it goes viral. When that happens who will stand up and take responsibility? It certainly won't be Mark Zuckerberg.
The inaccuracies generated by Meta AI are just the tip of the iceberg. On Monday the ABC's Media Watch drew attention to a sports news website it said was using AI to rip off and rewrite copy created by professional sports reporters.
And, surprise, surprise, the resulting content was as distorted, inaccurate and just plain wrong as the examples already discussed.
The AI conundrum is much bigger than just news reports. It was only last week the boss of the Australian Federal Police spoke about perverts using AI to generate child pornography. As a result police have wasted time and resources looking for victims who don't exist.
With the American election to be held this November and an Australian poll due in 2025 voters are going to be faced with a nightmare scenario.
Unless you are in the same room as Donald Trump, Joe Biden, Anthony Albanese or Peter Dutton how do you know what you are seeing and hearing is actually real?
If Meta AI has its way, the White Rabbit will lead us down into a dark and perverted "Wonderland" where all things are possible and nothing is real or true.
But the answer to that question of truth is as it has always been.
The best way to tell fact from fiction is to look to trusted sources. Credible media organisations staffed by professional journalists.
Credible journalism proved itself throughout the pandemic, when misinformation flourished on social media. People came to trusted news sites in droves when they and their communities most needed it.
Up against the explosion of fake news, real journalism has never been more important.
