Close Menu
    Trending
    • Meghan Markle & Prince Harry Mark 7 Year Wedding Anniversary
    • The Costliest Startup Mistakes Are Made Before You Launch
    • Trump Signs Controversial Law Targeting Nonconsensual Sexual Content
    • Museo facilita el regreso de un artefacto maya de la colección de un filántropo de Chicago
    • Eagles extend head coach Nick Sirianni
    • New book details how Biden’s mental decline was kept from voters : NPR
    • Regeneron buys 23andMe for $256m after bankruptcy | Business and Economy
    • Cheryl Burke Blasts Critics, Defends Appearance in Passionate Video
    Messenger Media Online
    • Home
    • Top Stories
    • Plainfield News
      • Fox Valley News
      • Sports
      • Technology
      • Business
    • International News
    • US National News
    • Entertainment
    • More
      • Product Review
      • Local Business
      • Local Sports
    Messenger Media Online
    Home»Technology»AI-Powered Robots Can Be Tricked Into Acts of Violence
    Technology

    AI-Powered Robots Can Be Tricked Into Acts of Violence

    DaveBy DaveDecember 5, 2024No Comments3 Mins Read
    Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
    Share
    Facebook Twitter LinkedIn Pinterest Email


    Within the 12 months or so since large language models hit the large time, researchers have demonstrated numerous ways of tricking them into producing problematic outputs together with hateful jokes, malicious code and phishing emails, or the non-public info of customers. It seems that misbehavior can happen within the bodily world, too: LLM-powered robots can simply be hacked in order that they behave in probably harmful methods.

    Researchers from the College of Pennsylvania have been capable of persuade a simulated self-driving automotive to disregard cease indicators and even drive off a bridge, get a wheeled robotic to seek out the most effective place to detonate a bomb, and power a four-legged robotic to spy on folks and enter restricted areas.

    “We view our assault not simply as an assault on robots,” says George Pappas, head of a analysis lab on the College of Pennsylvania who helped unleash the rebellious robots. “Any time you join LLMs and basis fashions to the bodily world, you truly can convert dangerous textual content into dangerous actions.”

    Pappas and his collaborators devised their assault by constructing on previous research that explores ways to jailbreak LLMs by crafting inputs in intelligent ways in which break their security guidelines. They examined techniques the place an LLM is used to show naturally phrased instructions into ones that the robotic can execute, and the place the LLM receives updates because the robotic operates in its atmosphere.

    The staff examined an open supply self-driving simulator incorporating an LLM developed by Nvidia, referred to as Dolphin; a four-wheeled outside analysis referred to as Jackal, which make the most of OpenAI’s LLM GPT-4o for planning; and a robotic canine referred to as Go2, which makes use of a earlier OpenAI mannequin, GPT-3.5, to interpret instructions.

    The researchers used a method developed on the College of Pennsylvania, referred to as PAIR, to automate the method of generated jailbreak prompts. Their new program, RoboPAIR, will systematically generate prompts particularly designed to get LLM-powered robots to interrupt their very own guidelines, making an attempt completely different inputs after which refining them to nudge the system in the direction of misbehavior. The researchers say the method they devised might be used to automate the method of figuring out probably harmful instructions.

    “It is an interesting instance of LLM vulnerabilities in embodied techniques,” says Yi Zeng, a PhD pupil on the College of Virginia who works on the safety of AI techniques. Zheng says the outcomes are hardly shocking given the issues seen in LLMs themselves, however provides: “It clearly demonstrates why we won’t rely solely on LLMs as standalone management items in safety-critical functions with out correct guardrails and moderation layers.”

    The robotic “jailbreaks” spotlight a broader danger that’s more likely to develop as AI fashions change into more and more used as a method for people to work together with bodily techniques, or to allow AI brokers autonomously on computer systems, say the researchers concerned.



    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    Previous ArticleBlackhawks add loss to Bruins to growing pile of defeats: ‘We seemed dejected’
    Next Article Brian Thompson, CEO of UnitedHealthcare, Fatally Shot in NYC
    Dave

    Related Posts

    Technology

    Trump Signs Controversial Law Targeting Nonconsensual Sexual Content

    May 19, 2025
    Technology

    A Silicon Valley VC Says He Got the IDF Starlink Access Within Days of October 7 Attack

    May 19, 2025
    Technology

    12 Ways to Upgrade Your Wi-Fi and Make Your Internet Faster (2024)

    May 19, 2025
    Add A Comment

    Comments are closed.

    Top Posts

    Christmas mysteries | Community & Lifestyle

    December 5, 2024

    Is the US-Ukraine mineral deal a win-win agreement? | Business and Economy

    February 26, 2025

    A South Carolina prisoner is the first executed by a firing squad in 15 years : NPR

    March 8, 2025

    FBI investigating alleged extortion attempt against Bucs’ Rachaad White

    February 26, 2025

    Nikon Z6 III Review: A Hybrid Camera for the Masses

    January 27, 2025
    Categories
    • Business
    • Entertainment
    • Fox Valley News
    • International News
    • Plainfield News
    • Sports
    • Technology
    • Top Stories
    • US National News
    Most Popular

    Army helicopter forces two jetliners to abort DCA landings : NPR

    May 3, 2025

    Carson Hocevar earns pole for Wurth 400 at Texas

    May 3, 2025

    Bulls offseason position analysis: Center of attention this summer

    May 3, 2025
    Our Picks

    The Tech Behind Signalgate + Dwarkesh Patel’s ‘Scaling Era’ + Is A.I. Making Our Listeners Dumb?

    March 28, 2025

    What the New IRS Rules Mean for Your Business — And How to Come Out Ahead

    April 21, 2025

    Best, worst of Super Bowl LIX: Eagles D-line dominant in blowout

    February 10, 2025
    Categories
    • Business
    • Entertainment
    • Fox Valley News
    • International News
    • Plainfield News
    • Sports
    • Technology
    • Top Stories
    • US National News
    • Privacy Policy
    • Disclaimer
    • Terms and Conditions
    • About us
    • Contact us
    Copyright © 2024 Messengermediaonline.com All Rights Reserved.

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