Pictures from your Instagram account could become fodder for Meta's new AI
Facebook parent company Meta has ditched its latest Artificial Intelligence (AI) feature after days of backlash.
Dubbed Muse Image, the tool was designed to let you generate AI images based on photos posted on any public Instagram account. And if your Instagram account was public, anybody could generate AI images with your likeness.
The feature was announced as a creative new way to connect with your friends and family.
For instance, if you wanted to generate a one-of-a-kind birthday card with a friend's face on it, you could tag their Instagram username and let the AI pull from public posts to start building the image. But critics said criminals could see this as a way to cook up convincing scams featuring your friends and family.
These types of scams are known as "deepfakes," relying on hyper-realistic digital manipulation of images, videos, or audio clips by AI to make someone appear to say or do something they never did. Deepfake scams have circulated on platforms like Instagram, WhatsApp, and Facebook for years. But the barrier to entry has always been fairly high.
After deciding to ditch Muse Image, Meta responded with a statement saying: "Earlier this week, we announced that one way for people to generate images in Meta AI is by @-mentioning public Instagram accounts that they want to reference.
"Our intent was to provide a useful creative tool and to give people control over whether their public content could be referenced in this way. We've heard the feedback that this feature missed the mark, so it's no longer available."
The introduction of Muse Image is a stark reminder that you're not always in control of what happens to your pictures or videos when you use these social media platforms. When the feature was initially released, it automatically opted in every public account. If you weren't aware of the arrival of Muse Image, you had unknowingly permitted others – well-intentioned or otherwise — to use your content.
Given that celebrities typically have public accounts, they would not have been immune. There have been several instances where fake accounts pretending to be public figures have tricked fans into handing over personal information. Taylor Swift was found to be the "most dangerous celebrity" in the UK late last year.
As AI becomes increasingly inserted into new updates across software and devices, it's not only getting harder to avoid. Apple has incorporated image generation capabilities via its Apple Intelligence system. Google has also integrated image generation via its AI assistant Gemini.
With Artificial Intelligence image-generation becoming widespread, social media users need to be on alert for these types of scams. Deepfakes are particularly dangerous because they exploit trust, convincing victims they're communicating with a family member or someone they know — making it more likely they'll send money or share personal information, like a password.
And it's not always possible to rely on the built-in safeguards developed by those who run the platforms to stop these scams. NBC News reported that Meta's own AI detection tools don't always catch these convincing fakes created from tools like Muse Image.
There have been multiple incidents over the years where bad actors from fake accounts or new numbers have reached out with an urgent request to get you to not think twice and hand over your information. We also almost fell victim to this type of scam when a fake mock account of a family member sent an urgent message of them being in immediate danger and needed help.
If you're unsure if you're looking at an AI-generated image, here are tips that may make it easier to spot:
And this isn't the first time Meta has come under fire regarding concerns over one's privacy.
Videos from Meta's Ray-Ban glasses of people undressing and using the toilet are being watched, a report claims.
High-resolution video filmed by the Meta-designed glasses is purportedly being reviewed by contractors at a Kenya-based company called Sama — a company responsible for training data that powers your computer vision and Artificial Intelligence (AI) technology.
Workers at the Nairobi facility say they've seen everything from people going to the toilet, getting undressed, and even having sex. According to the bombshell report, the footage arrives at their workstations largely unfiltered, meaning the Nairobi-based data annotators are watching deeply personal moments from glasses wearers.
The report has sparked outrage, with Will Wang, Even Realities' CEO, saying the rush to stick cameras on smart glasses is "really irresponsible."
Users who have purchased the glasses have also been found to be secretly recording other people, too.
Each pair of the Californian firm's smart glasses ships with a built-in camera and LED light. The latter flashes when you're taking a photo and keeps blinking while recording video. It's a critical privacy safeguard, so you're unable to capture photos and videos of your experiences hands-free without others knowing.
In order to bypass this, users have been hiring professionals to completely remove the light so they can record without notifying anyone.
