Carbon Chemist

Tag: photo sharing

  • This Website Shows How Much Google’s AI Can Glean From Your Photos

    This Website Shows How Much Google’s AI Can Glean From Your Photos

    [ad_1]

    Software engineer Vishnu Mohandas decided he would quit Google in more ways than one when he learned that the tech giant had briefly helped the US military develop AI to study drone footage. In 2020 he left his job working on Google Assistant and also stopped backing up all of his images to Google Photos. He feared that his content could be used to train AI systems, even if they weren’t specifically ones tied to the Pentagon project. “I don’t control any of the future outcomes that this will enable,” Mohandas thought. “So now, shouldn’t I be more responsible?”

    Mohandas, who taught himself programming and is based in Bengaluru, India, decided he wanted to develop an alternative service for storing and sharing photos that is open source and end-to-end encrypted. Something “more private, wholesome, and trustworthy,” he says. The paid service he designed, Ente, is profitable and says it has more than 100,000 users, many of whom are already part of the privacy-obsessed crowd. But Mohandas struggled to articulate to wider audiences why they should reconsider relying on Google Photos, despite all the conveniences it offers.

    Then one weekend in May, an intern at Ente came up with an idea: Give people a sense of what some of Google’s AI models can learn from studying images. Last month, Ente launched https://Theyseeyourphotos.com, a website and marketing stunt designed to turn Google’s technology against itself. People can upload any photo to the website, which is then sent to a Google Cloud computer vision program that writes a startlingly thorough three-paragraph description of it. (Ente prompts the AI model to document small details in the uploaded images.)

    One of the first photos Mohandas tried uploading was a selfie with his wife and daughter in front of a temple in Indonesia. Google’s analysis was exhaustive, even documenting the specific watch model that his wife was wearing, a Casio F-91W. But then, Mohandas says, the AI did something strange: It noted that Casio F-91W watches are commonly associated with Islamic extremists. “We had to tweak the prompts to make it slightly more wholesome but still spooky,” Mohandas says. Ente started asking the model to produce short, objective outputs—nothing dark.

    The same family photo uploaded to Theyseeyourphotos now returns a more generic result that includes the name of the temple and the “partly cloudy sky and lush greenery” surrounding it. But the AI still makes a number of assumptions about Mohandas and his family, like that their faces are expressing “joint contentment” and the “parents are likely of South Asian descent, middle class.” It judges their clothing (“appropriate for sightseeing”) and notes that “the woman’s watch displays a time as approximately 2 pm, which corroborates with the image metadata.”

    Google spokesperson Colin Smith declined to comment directly on Ente’s project. He directed WIRED to support pages that state uploads to Google Photos are only used to train generative AI models that help people manage their image libraries, like those that analyze the age and location of photo subjects.The company says it doesn’t sell the content stored in Google Photos to third parties or use it for advertising purposes. Users can turn off some of the analysis features in Photos, but they can’t prevent Google from accessing their images entirely, because the data are not end-to-end encrypted.

    [ad_2]

    Source link

  • Photography Is No Longer Evidence of Anything

    Photography Is No Longer Evidence of Anything

    [ad_1]

    For weeks now, the world has been awash in conspiracy theories spurred by weird artifacts in a photographic image of the missing Princess of Wales that she eventually admitted had been edited. Some of them got pretty crazy, ranging from a cover-up of Kate’s alleged death, to a theory that the Royal Family were reptilian aliens. But none was as bizarre as the idea that in 2024 anyone might believe that a digital image is evidence of anything.

    Not only are digital images infinitely malleable, but the tools to manipulate them are as common as dirt. For anyone paying attention, this has been clear for decades. The issue was definitively laid out almost 40 years ago, in a piece cowritten by Kevin Kelly, a founding WIRED editor; Stewart Brand; and Jay Kinney in the July 1985 edition of The Whole Earth Review, a publication run out of Brand’s organization in Sausalito, California. Kelly had gotten the idea for the story a year or so earlier when he came across an internal newsletter for publisher Time Life, where his father worked. It described a million-dollar machine called Scitex, which created high-resolution digital images from photographic film, which could then be altered using a computer. High-end magazines were among the first customers: Kelly learned that National Geographic had used the tool to literally move one of the Pyramids of Giza so it could fit into a cover shot. “I thought, ‘Man, this is gonna change everything,’” says Kelly.

    The article was titled “Digital Retouching: The End of Photography as Evidence of Anything.” It opened with an imaginary courtroom scene where a lawyer argued that compromising photos should be excluded from a case, saying that due to its unreliability, “photography has no place in this or any other courtroom. For that matter, neither does film, videotape, or audiotape.”

    Did the article draw wide attention to the fact that photography might be stripped of its role as documentary proof, or the prospect of an era where no one can tell what’s real or fake? “No!” says Kelly. No one noticed. Even Kelly thought it would be many years before the tools to convincingly alter photos would become routinely available. Three years later, two brothers from Michigan invented what would become Photoshop, released as an Adobe product in 1990. The application put digital photo manipulation on desktop PCs, cutting the cost dramatically. By then even The New York Times was reporting on “the ethical issues involved in altering photographs and other materials using digital editing.”

    Adobe, in the eye of this storm for decades, has given a lot of thought to those issues. Ely Greenfield, CTO of Adobe’s digital media business, rightfully points out that long before Photoshop, film photographers and cinematographers used tricks to alter their images. But even though digital tools make the practice cheap and commonplace, Greenfield says, “treating photos and videos as documentary sources of truth is still a valuable thing. What is the purpose of an image? Is it there to look pretty? Is it there to tell a story? We all like looking at pretty images. But we think there’s still value in the storytelling.”

    To ascertain whether photographic storytelling is accurate or faked, Adobe and others have devised a tool set that strives for a degree of verifiability. Metadata in the Middleton photo, for instance, helped people ascertain that its anomalies were the result of a Photoshop edit, which the Princess owned up to. A consortium of over 2,500 creators, technologists, and publishers called the Content Authenticity Initiative, started by Adobe in 2019, is working to devise tools and standards so people can verify whether an image, video, or recording has been altered. It’s based on combining metadata with exotic watermarking and cryptographic techniques. Greenfield concedes, though, that those protections can be circumvented. “We have technologies that can detect edited photos or AI-generated photos, but it’s still a losing battle,” he says. “As long as there is a motivated enough actor who’s determined to overcome those technologies, they will.”



    [ad_2]

    Source link