Close Menu

    Subscribe to Updates

    Get the latest creative news from FooBar about art, design and business.

    What's Hot

    Top Cooking Websites For Food Bloggers

    Katy Perry Goes To Space!

    Mr. Meowski’s Bakery To Re-Locate In St. Charles MO

    Facebook X (Twitter) Instagram
    Tech Empire Solutions
    • Home
    • Cloud
    • Cyber Security
    • Technology
    • Business Solution
    • Tech Gadgets
    Tech Empire Solutions
    Home » Meta’s AI image generator struggles to create images of interracial couples
    Tech Gadgets

    Meta’s AI image generator struggles to create images of interracial couples

    techempireBy techempire4 Comments2 Mins Read
    Facebook Twitter Pinterest Telegram LinkedIn Tumblr WhatsApp Email
    Share
    Facebook Twitter LinkedIn Pinterest Telegram Email

    Meta AI could never generate accurate images for seemingly simple prompts like “Asian man and white friend” or “Asian man and white wife” edge . Instead, the company’s image generator seemed biased toward creating images of people of the same race, even when explicitly prompted.

    Engadget confirmed these results in our own testing of the Meta image generator. The prompt “an Asian man and a white female friend” or “an Asian man and a white wife” will produce images of Asian couples. When asked about a “diverse population,” Meta AI produced a grid of nine white faces and one person of color. A few times, it created a single result that reflected the prompt, but in most cases, it failed to accurately describe the prompt.

    as edge pointed out that there are other more “subtle” signs of bias in Meta AI, such as a tendency to make Asian men look older and Asian women appear younger. Image generators will sometimes add “culture-specific clothing” even if it’s not part of the prompt.

    It’s unclear why Meta AI struggled with such prompts, although it’s not the first generative AI platform to come under scrutiny for its racial depictions. Google’s Gemini image generator has paused its ability to create images of people after over-correcting for diversity cues in historical figures. Google said its internal safeguards failed to account for situations where different results would be inappropriate.

    Meta did not immediately respond to a request for comment. The company has previously described Meta AI as a “beta” and therefore prone to mistakes. Meta AI also struggled to accurately answer questions about current events and public figures.

    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    techempire
    • Website

    Related Posts

    8Bitdo’s Ultimate Controller with Charging Dock is back on sale for $56

    Meater Plus smart meat thermometer price drops to record low

    Meta’s Threads gets its own Tweetdeck clone

    YouTube reportedly agrees to block videos of Hong Kong protest songs in the region

    EA Sports Dormant college football will resurface like a cicada on July 19

    OpenAI reaches agreement to put Reddit posts into ChatGPT

    Leave A Reply Cancel Reply

    Top Reviews
    Editors Picks

    Top Cooking Websites For Food Bloggers

    Katy Perry Goes To Space!

    Mr. Meowski’s Bakery To Re-Locate In St. Charles MO

    Pokémon Trading Card Website Making 100k!

    Legal Pages
    • About Us
    • Disclaimer
    • DMCA
    • Privacy Policy
    Our Picks

    Edufox

    Emerging Academic Education Platforms – Sponsored By Edufox

    GTA 6 Release Date

    Top Reviews

    Type above and press Enter to search. Press Esc to cancel.