That AI clip of Tom Cruise and Brad Pitt fighting might just have accelerated YouTube’s response to deepfakes and AI. You may well have already noticed this, but if you’re doing a search for something relatively contemporary on YouTube at the moment, the service is absolutely awash with AI-generated material. With no clear way to ... YouTube introduced tool to identify AI deepfakes (but only for rich people)
That AI clip of Tom Cruise and Brad Pitt fighting might just have accelerated YouTube’s response to deepfakes and AI.
You may well have already noticed this, but if you’re doing a search for something relatively contemporary on YouTube at the moment, the service is absolutely awash with AI-generated material. With no clear way to exclude AI material from search results, it’s making browsing the YouTube website more of a challenge than it should be.
It’s not just boring people living in Halesowen in the UK that are bothered by this. Powerful people who actually have money have something to say about it too.
Headlines were made a few months ago, after all, when a deepfake sequence involving Brad Pitt and Tom Cruise having a scrap went viral around the world. It was material, inevitably, where an AI system had harvested previous footage and represented it, but it certainly – and understandably – made a few meetings happen around Hollywood.
Perhaps in response to that, YouTube has acted. It’s been working on a deepfake detection tool, which is going to be available to those at risk of having their likeness misused. Famous and rich people, basically.
They actually have to sign up to the service, which will then identify deepfakes of them on YouTube, and allow them to request their takedown.
Mary Ellen Coe is chief business officer at YouTube, and has been nattering to The Hollywood Reporter about the tool, which she says has been in testing for around 18 months. It’s thus far been available to ‘selected’ politicans, and just because someone complains about a deepfake, it turns out YouTube won’t necessarily take it down. “There are a number of cases like parody and satire where our community guidelines would allow that to remain on the platform”, Coe said.
Instead, “If someone is doing an exact replica of something that would limit the livelihood of a celebrity, an actor or a creator, because it’s literal content replacement, that would be included in a takedown.”
Sounds a bit of a muddle, this, although with hugely expensive lawsuits around AI, it feels as if YouTube had to do something, before it was a) made to, or b) on the receiving end of a massive bill.
Separately, YouTube’s owner, Google, continues to use its AI services to scrape material of websites such as this, and present it in its AI search results. This, thus, denies websites a click and any financial reward for their work, and Google provides no recompense in return. But hey, we’re not Brad Pitt or Tom Cruise, so no tool has been provided for the likes of us. Just in case you’re wondering why so many editorially-driven websites are being decimated at the moment.
Might get a coffee.
