Close Menu
Global News HQ
    What's Hot

    It’s Official: These Are The Best Running Leggings *Ever* — Here’s Why

    November 16, 2025

    British Columbia Employers: New Sick Note Rules Now in Effect

    November 16, 2025

    Crypto’s long-term fundamentals strong despite market rout — Bitwise CEO

    November 16, 2025
    Recent Posts
    • It’s Official: These Are The Best Running Leggings *Ever* — Here’s Why
    • British Columbia Employers: New Sick Note Rules Now in Effect
    • Crypto’s long-term fundamentals strong despite market rout — Bitwise CEO
    • 6 Things Minimalists Always Do for the Tidiest Kitchen, According to Pro Organizers
    • Medicare Premiums 2026: IRMAA Brackets and Surcharges for Parts B and D
    Facebook X (Twitter) Instagram YouTube TikTok
    Trending
    • It’s Official: These Are The Best Running Leggings *Ever* — Here’s Why
    • British Columbia Employers: New Sick Note Rules Now in Effect
    • Crypto’s long-term fundamentals strong despite market rout — Bitwise CEO
    • 6 Things Minimalists Always Do for the Tidiest Kitchen, According to Pro Organizers
    • Medicare Premiums 2026: IRMAA Brackets and Surcharges for Parts B and D
    • Get Started in Social Media Marketing With This Step-By-Step Guide
    • What you need to know about Sonder if you have a canceled or upcoming reservation – The Points Guy
    • Apple is reportedly getting ready to replace Tim Cook as early as next year
    Global News HQ
    • Technology & Gadgets
    • Travel & Tourism (Luxury)
    • Health & Wellness (Specialized)
    • Home Improvement & Remodeling
    • Luxury Goods & Services
    • Home
    • Finance & Investment
    • Insurance
    • Legal
    • Real Estate
    • More
      • Cryptocurrency & Blockchain
      • E-commerce & Retail
      • Business & Entrepreneurship
      • Automotive (Car Deals & Maintenance)
    Global News HQ
    Home - Technology & Gadgets - Researchers surprised that with AI, toxicity is harder to fake than intelligence
    Technology & Gadgets

    Researchers surprised that with AI, toxicity is harder to fake than intelligence

    Facebook Twitter Pinterest LinkedIn Tumblr WhatsApp VKontakte Email
    Researchers surprised that with AI, toxicity is harder to fake than intelligence
    Share
    Facebook Twitter LinkedIn Pinterest Email



    The next time you encounter an unusually polite reply on social media, you might want to check twice. It could be an AI model trying (and failing) to blend in with the crowd.

    On Wednesday, researchers from the University of Zurich, University of Amsterdam, Duke University, and NYU released a study revealing that AI models remain easily distinguishable from humans in social media conversations, with overly friendly emotional tone serving as the most persistent giveaway. The research, which tested nine open-weight models across Twitter/X, Bluesky, and Reddit, found that classifiers developed by the researchers detected AI-generated replies with 70 to 80 percent accuracy.

    The study introduces what the authors call a “computational Turing test” to assess how closely AI models approximate human language. Instead of relying on subjective human judgment about whether text sounds authentic, the framework uses automated classifiers and linguistic analysis to identify specific features that distinguish machine-generated from human-authored content.

    “Even after calibration, LLM outputs remain clearly distinguishable from human text, particularly in affective tone and emotional expression,” the researchers wrote. The team, led by Nicolò Pagan at the University of Zurich, tested various optimization strategies, from simple prompting to fine-tuning, but found that deeper emotional cues persist as reliable tells that a particular text interaction online was authored by an AI chatbot rather than a human.

    The toxicity tell

    In the study, researchers tested nine large language models: Llama 3.1 8B, Llama 3.1 8B Instruct, Llama 3.1 70B, Mistral 7B v0.1, Mistral 7B Instruct v0.2, Qwen 2.5 7B Instruct, Gemma 3 4B Instruct, DeepSeek-R1-Distill-Llama-8B, and Apertus-8B-2509.

    When prompted to generate replies to real social media posts from actual users, the AI models struggled to match the level of casual negativity and spontaneous emotional expression common in human social media posts, with toxicity scores consistently lower than authentic human replies across all three platforms.

    To counter this deficiency, the researchers attempted optimization strategies (including providing writing examples and context retrieval) that reduced structural differences like sentence length or word count, but variations in emotional tone persisted. “Our comprehensive calibration tests challenge the assumption that more sophisticated optimization necessarily yields more human-like output,” the researchers concluded.



    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr WhatsApp Email
    Previous ArticleAudemars Piguet goes global with new boutique concept
    Next Article American Express rolls out experiential series at F1 Las Vegas Grand Prix

    Related Posts

    Apple is reportedly getting ready to replace Tim Cook as early as next year

    November 15, 2025

    6 Vitamins and Minerals for Older Adults, According to the Experts

    November 15, 2025

    Wyoming dinosaur mummies give us a new view of duck-billed species

    November 15, 2025

    YouTube TV, ESPN, and Disney: the latest on the blackout that’s now over

    November 15, 2025
    Leave A Reply Cancel Reply

    ads
    Don't Miss
    Health & Wellness (Specialized)
    15 Mins Read

    It’s Official: These Are The Best Running Leggings *Ever* — Here’s Why

    Bold statement alert: Leggings can make or break a run. Tights that slide down the…

    British Columbia Employers: New Sick Note Rules Now in Effect

    November 16, 2025

    Crypto’s long-term fundamentals strong despite market rout — Bitwise CEO

    November 16, 2025

    6 Things Minimalists Always Do for the Tidiest Kitchen, According to Pro Organizers

    November 16, 2025
    Top
    Health & Wellness (Specialized)
    15 Mins Read

    It’s Official: These Are The Best Running Leggings *Ever* — Here’s Why

    Bold statement alert: Leggings can make or break a run. Tights that slide down the…

    British Columbia Employers: New Sick Note Rules Now in Effect

    November 16, 2025

    Crypto’s long-term fundamentals strong despite market rout — Bitwise CEO

    November 16, 2025
    Our Picks
    Health & Wellness (Specialized)
    15 Mins Read

    It’s Official: These Are The Best Running Leggings *Ever* — Here’s Why

    Bold statement alert: Leggings can make or break a run. Tights that slide down the…

    Legal
    1 Min Read

    British Columbia Employers: New Sick Note Rules Now in Effect

    British Columbia Employers: New Sick Note Rules Now in Effect Source link

    Pages
    • About Us
    • Contact Us
    • Disclaimer
    • Homepage
    • Privacy Policy
    Facebook X (Twitter) Instagram YouTube TikTok
    • Home
    © 2025 Global News HQ .

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

    Go to mobile version