Close Menu
Must Have Gadgets –

    Subscribe to Updates

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

    What's Hot

    I test audio gear for a living and this is the Dolby Atmos soundbar I recommend above everything else

    November 8, 2025

    Branch Sale of the Year: Deals on Office Chairs, Standing Desks, and Home Office Gear

    November 8, 2025

    LegalZoom Promo Code: Exclusive 10% Off LLC Formations

    November 8, 2025
    Facebook X (Twitter) Instagram
    Must Have Gadgets –
    Trending
    • I test audio gear for a living and this is the Dolby Atmos soundbar I recommend above everything else
    • Branch Sale of the Year: Deals on Office Chairs, Standing Desks, and Home Office Gear
    • LegalZoom Promo Code: Exclusive 10% Off LLC Formations
    • Paramount Plus Coupon Codes and Deals: Up to 50% Off
    • Big Tech’s tax breaks could have covered benefits for millions of Americans
    • Fitbit’s food logging is down, but that’s no excuse for a cheat day
    • This Early Black Friday Deal Brings the Galaxy S25 Ultra Down to a New Record Low Price, but Not for Long
    • GTA 6 Has Been Delayed Once Again — Prepare to Wait
    • Home
    • Shop
      • Earbuds & Headphones
      • Smartwatches
      • Mobile Accessories
      • Smart Home Devices
      • Laptops & Tablets
    • Gadget Reviews
    • How-To Guides
    • Mobile Accessories
    • Smart Devices
    • More
      • Top Deals
      • Smart Home
      • Tech News
      • Trending Tech
    Facebook X (Twitter) Instagram
    Must Have Gadgets –
    Home»Tech News»Researchers surprised that with AI, toxicity is harder to fake than intelligence
    Tech News

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

    adminBy adminNovember 7, 2025No Comments2 Mins Read
    Facebook Twitter Pinterest LinkedIn Tumblr 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 New York University 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.

    Fake harder intelligence Researchers surprised toxicity
    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    admin
    • Website

    Related Posts

    Paramount Plus Coupon Codes and Deals: Up to 50% Off

    November 8, 2025

    Texas sues Roblox for allegedly failing to protect children on its platform

    November 8, 2025

    The redesigned Disney+ app is rolling out to more users in the US

    November 8, 2025
    Leave A Reply Cancel Reply

    Top Posts

    I test audio gear for a living and this is the Dolby Atmos soundbar I recommend above everything else

    November 8, 2025

    PayPal’s blockchain partner accidentally minted $300 trillion in stablecoins

    October 16, 2025

    The best AirPods deals for October 2025

    October 16, 2025
    Stay In Touch
    • Facebook
    • YouTube
    • TikTok
    • WhatsApp
    • Twitter
    • Instagram
    Latest Reviews
    How-To Guides

    How to Disable Some or All AI Features on your Samsung Galaxy Phone

    By adminOctober 16, 20250
    Gadget Reviews

    PayPal’s blockchain partner accidentally minted $300 trillion in stablecoins

    By adminOctober 16, 20250
    Smart Devices

    The best AirPods deals for October 2025

    By adminOctober 16, 20250

    Subscribe to Updates

    Get the latest tech news from FooBar about tech, design and biz.

    Latest Post

    I test audio gear for a living and this is the Dolby Atmos soundbar I recommend above everything else

    November 8, 2025

    Branch Sale of the Year: Deals on Office Chairs, Standing Desks, and Home Office Gear

    November 8, 2025

    LegalZoom Promo Code: Exclusive 10% Off LLC Formations

    November 8, 2025
    Recent Posts
    • I test audio gear for a living and this is the Dolby Atmos soundbar I recommend above everything else
    • Branch Sale of the Year: Deals on Office Chairs, Standing Desks, and Home Office Gear
    • LegalZoom Promo Code: Exclusive 10% Off LLC Formations
    • Paramount Plus Coupon Codes and Deals: Up to 50% Off
    • Big Tech’s tax breaks could have covered benefits for millions of Americans

    Subscribe to Updates

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

    Facebook
    • About Us
    • Contact Us
    • Privacy Policy
    • Terms and Conditions
    • Disclaimer
    © 2025 must-have-gadgets.

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