An AI-generated image also depicted what appeared to be French President Emmanuel Macron embroiled in a riot.
AI-powered technology is helping generate fake images, videos and audio that are becoming increasingly difficult to distinguish from reality.
There are five clues to finding AI-generated images. These include focusing on hands, background images, and inanimate objects.
As technology evolves, we must learn how to navigate new challenges. Acquiring AI detective skills empowers us. Trump’s fake skepticism shows that we’re not AI suckers for her. (A Manhattan grand jury voted to indict Trump on Thursday.)
But tell the truth. I feel uneasy writing this newsletter. I don’t want to hype your concerns about AI fakes, which are themselves risky. A focus on AI forensics can also distract from the deeper reason that fakes are attractive.
15 years of social media history and centuries of conspiracy theories show that the sophistication of the “evidence” makes false information unbelievable. We fall for the fakes when we want to believe the reality they present.
5 hints that the image may be fake, generated by AI
1. Look at your hands. AI software has a history of generating human hands with quirky hands, such as having too many fingers. Although this technology is now becoming viable, it still has many problems.
For example, in the fake image of Pope Francis, his right hand appears to be squashed and what appears to be the coffee cup he appears to be holding. You can see that it is not.
2. Inanimate objects may be off-base. AI software, including Midjourney’s AI software used to create the puffy Pope’s coat and fake Trump arrest images, can generate objects that defy reality.
To find it, focus on items in the image such as glasses, fences, and bicycles.
Some keen-eyed people have noticed that the fake image of Pope Francis has only one strap on the traditional pectoral cross around his neck.
The computer-generated person may have missing earrings or mismatched eyepiece earpieces. These flaws were more pronounced in previous generations of his AI imaging software, but these distortions still appear.
Machines can also be unwieldy for AI. Journalist Luke Bailey An AI-generated unicycle that was laughably off base.
3. Are the characters garbled? If you’re wondering if an image was created by AI, look for what’s written on objects such as road signs and billboards.
Bailey also showed an AI-generated image of Prince Harry holding a bag of McDonald’s food. The logo for his chain of restaurants looked real, but the text on the bag was gibberish.
4. Scan the background. AI-generated images can have blurry or distorted details, especially in the background.
In one of Trump’s fake images, the law enforcement officer’s face appeared blurry or disfigured. In another example, the eyes of an AI-generated fake police officer appeared to point in the wrong direction.
5. Does the image look too glossy or artistic? Some AI-generated images of real people are gaudy stylized, or depict people with faces that look like plastic.
The AI-generated face of Pope Francis had an “aesthetic sheen,” said Henry Ajder, an expert in manipulated or artificially generated media. “The AI software smoothes them out a little too much, making them too shiny.”
As AI technology advances, it becomes more difficult to find AI-generated people and altered images. Ajder warned that clues to identifying AI images could quickly become obsolete. “In a matter of weeks, we can train these defects from these models,” he said.
The Big Picture: Treating AI Fakes Like Apocalypse Is Irresponsible
Fake images are nothing new. For example, over the past decade, fake images of sharks purportedly swimming through flooded city streets during hurricanes and other storms have circulated repeatedly.
But it’s scary how AI software gives most people the ability to create compelling-looking images in minutes.
Our challenge is to deal with the risk of AI counterfeiting without being too small or worrying enough to cause self-contained panic.
Researchers talk about a phenomenon known as the “liar’s dividend.” The more we believe that what we see and hear is fake, the more we risk discrediting the veracity of everything. This Orwellian distrust is what authoritarian governments like. You and I have to resist this.
It’s also important to recognize that fakes and hoaxes are part of our lives forever.
I’m also addicted to fakes. Early in the 2020 coronavirus outbreak, I saw a viral tweet containing an image of Tom Hanks supposedly locked in a hospital room with Wilson, the volleyball player in the movie Cast Away.
The image was a photoshopped fake from a satirical Australian news publication, but I retweeted it without a second thought. I feared the pandemic, but this lighthearted moment felt like a relief. i wanted it to be true.
Claire Wardle, co-founder of Brown University’s Information Futures Lab, said she was encouraged that relatively few people believed the AI-generated image of Trump was real.
She said it shows that many of us are learning to identify and seek confirmation of what we see online. He said he had seen comments on Twitter from people who said the information would have been posted on traditional news sites.
“It’s easy to go down the doomsday road, but I think we’re actually smarter than we think we are,” Wardle said.
One of the biggest headaches for help desk readers is having their Facebook accounts taken over by hackers.
Also, Facebook stinks about making account recovery easy. My colleague Heather Kelly has a suggestion on how to avoid hijacking her Facebook account in the first place.
Turn on two-factor authentication when you only do one thing – Password plus extra steps like a secret code to access your Facebook account. do this:
Tap the three lines in the top right corner (Android app) or the three lines in the bottom right corner (iPhone app) → Scroll down to[設定とプライバシー]→[設定]→ at the top of the screen[Meta Accounts Center]→[パスワードとセキュリティ]→ Two-Factor Authentication → Click Edit and enter your Facebook password.
You can choose from three options. Most people should choose one of these two for her.
- Text message (SMS): After you enter your password, Facebook will text your phone a number that you must enter on the website or Facebook app to log in.
- Authenticator app: This works like the text option, but opens a third-party app to get a numeric code instead of a message. We recommend Twilio’s Authy or Google Authenticator (iOS, Android).
Tired of me recommending two-factor authentication? I’m going to keep doing it until our entire ridiculous password system is shattered.
- Don’t get hacked on Facebook. Do these 6 things now.
- Everything said about passwords is a lie