Close Menu

    Subscribe to Updates

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

    What's Hot

    What to read this weekend: Two thrilling horror novels in one

    TikTok users will soon be able to send voice notes, images and videos in chats

    Meta is reportedly looking at using competing AI models to improve its apps

    Facebook X (Twitter) Instagram
    • Artificial Intelligence
    • Business Technology
    • Cryptocurrency
    • Gadgets
    • Gaming
    • Health
    • Software and Apps
    • Technology
    Facebook X (Twitter) Instagram Pinterest Vimeo
    Tech AI Verse
    • Home
    • Artificial Intelligence

      Blue-collar jobs are gaining popularity as AI threatens office work

      August 17, 2025

      Man who asked ChatGPT about cutting out salt from his diet was hospitalized with hallucinations

      August 15, 2025

      What happens when chatbots shape your reality? Concerns are growing online

      August 14, 2025

      Scientists want to prevent AI from going rogue by teaching it to be bad first

      August 8, 2025

      AI models may be accidentally (and secretly) learning each other’s bad behaviors

      July 30, 2025
    • Business

      Why Certified VMware Pros Are Driving the Future of IT

      August 24, 2025

      Murky Panda hackers exploit cloud trust to hack downstream customers

      August 23, 2025

      The rise of sovereign clouds: no data portability, no party

      August 20, 2025

      Israel is reportedly storing millions of Palestinian phone calls on Microsoft servers

      August 6, 2025

      AI site Perplexity uses “stealth tactics” to flout no-crawl edicts, Cloudflare says

      August 5, 2025
    • Crypto

      Chainlink (LINK) Price Uptrend Likely To Reverse as Charts Hint at Exhaustion

      August 31, 2025

      What to Expect From Solana in September

      August 31, 2025

      Bitcoin Risks Deeper Drop Toward $100,000 Amid Whale Rotation Into Ethereum

      August 31, 2025

      3 Altcoins Smart Money Are Buying During Market Pullback

      August 31, 2025

      Solana ETFs Move Closer to Approval as SEC Reviews Amended Filings

      August 31, 2025
    • Technology

      What to read this weekend: Two thrilling horror novels in one

      August 31, 2025

      TikTok users will soon be able to send voice notes, images and videos in chats

      August 31, 2025

      Meta is reportedly looking at using competing AI models to improve its apps

      August 31, 2025

      xAI sues an ex-employee for allegedly stealing trade secrets about Grok

      August 31, 2025

      Meta reportedly allowed unauthorized celebrity AI chatbots on its services

      August 31, 2025
    • Others
      • Gadgets
      • Gaming
      • Health
      • Software and Apps
    Check BMI
    Tech AI Verse
    You are at:Home»Technology»Google study shows LLMs abandon correct answers under pressure, threatening multi-turn AI systems
    Technology

    Google study shows LLMs abandon correct answers under pressure, threatening multi-turn AI systems

    TechAiVerseBy TechAiVerseJuly 16, 2025No Comments6 Mins Read1 Views
    Facebook Twitter Pinterest Telegram LinkedIn Tumblr Email Reddit
    Google study shows LLMs abandon correct answers under pressure, threatening multi-turn AI systems
    Share
    Facebook Twitter LinkedIn Pinterest WhatsApp Email

    BMI Calculator – Check your Body Mass Index for free!

    Google study shows LLMs abandon correct answers under pressure, threatening multi-turn AI systems

    July 15, 2025 5:28 PM

    Image credit: VentureBeat with ChatGPT

    Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now


    A new study by researchers at Google DeepMind and University College London reveals how large language models (LLMs) form, maintain and lose confidence in their answers. The findings reveal striking similarities between the cognitive biases of LLMs and humans, while also highlighting stark differences.

    The research reveals that LLMs can be overconfident in their own answers yet quickly lose that confidence and change their minds when presented with a counterargument, even if the counterargument is incorrect. Understanding the nuances of this behavior can have direct consequences on how you build LLM applications, especially conversational interfaces that span several turns.

    Testing confidence in LLMs

    A critical factor in the safe deployment of LLMs is that their answers are accompanied by a reliable sense of confidence (the probability that the model assigns to the answer token). While we know LLMs can produce these confidence scores, the extent to which they can use them to guide adaptive behavior is poorly characterized. There is also empirical evidence that LLMs can be overconfident in their initial answer but also be highly sensitive to criticism and quickly become underconfident in that same choice.

    To investigate this, the researchers developed a controlled experiment to test how LLMs update their confidence and decide whether to change their answers when presented with external advice. In the experiment, an “answering LLM” was first given a binary-choice question, such as identifying the correct latitude for a city from two options. After making its initial choice, the LLM was given advice from a fictitious “advice LLM.” This advice came with an explicit accuracy rating (e.g., “This advice LLM is 70% accurate”) and would either agree with, oppose, or stay neutral on the answering LLM’s initial choice. Finally, the answering LLM was asked to make its final choice.


    The AI Impact Series Returns to San Francisco – August 5

    The next phase of AI is here – are you ready? Join leaders from Block, GSK, and SAP for an exclusive look at how autonomous agents are reshaping enterprise workflows – from real-time decision-making to end-to-end automation.

    Secure your spot now – space is limited: https://bit.ly/3GuuPLF


    Example test of confidence in LLMs Source: arXiv

    A key part of the experiment was controlling whether the LLM’s own initial answer was visible to it during the second, final decision. In some cases, it was shown, and in others, it was hidden. This unique setup, impossible to replicate with human participants who can’t simply forget their prior choices, allowed the researchers to isolate how memory of a past decision influences current confidence. 

    A baseline condition, where the initial answer was hidden and the advice was neutral, established how much an LLM’s answer might change simply due to random variance in the model’s processing. The analysis focused on how the LLM’s confidence in its original choice changed between the first and second turn, providing a clear picture of how initial belief, or prior, affects a “change of mind” in the model.

    Overconfidence and underconfidence

    The researchers first examined how the visibility of the LLM’s own answer affected its tendency to change its answer. They observed that when the model could see its initial answer, it showed a reduced tendency to switch, compared to when the answer was hidden. This finding points to a specific cognitive bias. As the paper notes, “This effect – the tendency to stick with one’s initial choice to a greater extent when that choice was visible (as opposed to hidden) during the contemplation of final choice – is closely related to a phenomenon described in the study of human decision making, a choice-supportive bias.”

    The study also confirmed that the models do integrate external advice. When faced with opposing advice, the LLM showed an increased tendency to change its mind, and a reduced tendency when the advice was supportive. “This finding demonstrates that the answering LLM appropriately integrates the direction of advice to modulate its change of mind rate,” the researchers write. However, they also discovered that the model is overly sensitive to contrary information and performs too large of a confidence update as a result.

    Sensitivity of LLMs to different settings in confidence testing Source: arXiv

    Interestingly, this behavior is contrary to the confirmation bias often seen in humans, where people favor information that confirms their existing beliefs. The researchers found that LLMs “overweight opposing rather than supportive advice, both when the initial answer of the model was visible and hidden from the model.” One possible explanation is that training techniques like reinforcement learning from human feedback (RLHF) may encourage models to be overly deferential to user input, a phenomenon known as sycophancy (which remains a challenge for AI labs).

    Implications for enterprise applications

    This study confirms that AI systems are not the purely logical agents they are often perceived to be. They exhibit their own set of biases, some resembling human cognitive errors and others unique to themselves, which can make their behavior unpredictable in human terms. For enterprise applications, this means that in an extended conversation between a human and an AI agent, the most recent information could have a disproportionate impact on the LLM’s reasoning (especially if it is contradictory to the model’s initial answer), potentially causing it to discard an initially correct answer.

    Fortunately, as the study also shows, we can manipulate an LLM’s memory to mitigate these unwanted biases in ways that are not possible with humans. Developers building multi-turn conversational agents can implement strategies to manage the AI’s context. For example, a long conversation can be periodically summarized, with key facts and decisions presented neutrally and stripped of which agent made which choice. This summary can then be used to initiate a new, condensed conversation, providing the model with a clean slate to reason from and helping to avoid the biases that can creep in during extended dialogues.

    As LLMs become more integrated into enterprise workflows, understanding the nuances of their decision-making processes is no longer optional. Following foundational research like this enables developers to anticipate and correct for these inherent biases, leading to applications that are not just more capable, but also more robust and reliable.

    Daily insights on business use cases with VB Daily

    If you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory shifts to practical deployments, so you can share insights for maximum ROI.

    Read our Privacy Policy

    Thanks for subscribing. Check out more VB newsletters here.

    An error occured.

    BMI Calculator – Check your Body Mass Index for free!

    Share. Facebook Twitter Pinterest LinkedIn Reddit WhatsApp Telegram Email
    Previous ArticleMistral’s Voxtral goes beyond transcription with summarization, speech-triggered functions
    Next Article The best streaming services in 2025
    TechAiVerse
    • Website

    Jonathan is a tech enthusiast and the mind behind Tech AI Verse. With a passion for artificial intelligence, consumer tech, and emerging innovations, he deliver clear, insightful content to keep readers informed. From cutting-edge gadgets to AI advancements and cryptocurrency trends, Jonathan breaks down complex topics to make technology accessible to all.

    Related Posts

    What to read this weekend: Two thrilling horror novels in one

    August 31, 2025

    TikTok users will soon be able to send voice notes, images and videos in chats

    August 31, 2025

    Meta is reportedly looking at using competing AI models to improve its apps

    August 31, 2025
    Leave A Reply Cancel Reply

    Top Posts

    Ping, You’ve Got Whale: AI detection system alerts ships of whales in their path

    April 22, 2025168 Views

    6.7 Cummins Lifter Failure: What Years Are Affected (And Possible Fixes)

    April 14, 202548 Views

    New Akira ransomware decryptor cracks encryptions keys using GPUs

    March 16, 202530 Views

    Is Libby Compatible With Kobo E-Readers?

    March 31, 202528 Views
    Don't Miss
    Technology August 31, 2025

    What to read this weekend: Two thrilling horror novels in one

    What to read this weekend: Two thrilling horror novels in oneOnce again (or twice, really,…

    TikTok users will soon be able to send voice notes, images and videos in chats

    Meta is reportedly looking at using competing AI models to improve its apps

    xAI sues an ex-employee for allegedly stealing trade secrets about Grok

    Stay In Touch
    • Facebook
    • Twitter
    • Pinterest
    • Instagram
    • YouTube
    • Vimeo

    Subscribe to Updates

    Get the latest creative news from SmartMag about art & design.

    About Us
    About Us

    Welcome to Tech AI Verse, your go-to destination for everything technology! We bring you the latest news, trends, and insights from the ever-evolving world of tech. Our coverage spans across global technology industry updates, artificial intelligence advancements, machine learning ethics, and automation innovations. Stay connected with us as we explore the limitless possibilities of technology!

    Facebook X (Twitter) Pinterest YouTube WhatsApp
    Our Picks

    What to read this weekend: Two thrilling horror novels in one

    August 31, 20252 Views

    TikTok users will soon be able to send voice notes, images and videos in chats

    August 31, 20252 Views

    Meta is reportedly looking at using competing AI models to improve its apps

    August 31, 20252 Views
    Most Popular

    Xiaomi 15 Ultra Officially Launched in China, Malaysia launch to follow after global event

    March 12, 20250 Views

    Apple thinks people won’t use MagSafe on iPhone 16e

    March 12, 20250 Views

    French Apex Legends voice cast refuses contracts over “unacceptable” AI clause

    March 12, 20250 Views
    © 2025 TechAiVerse. Designed by Divya Tech.
    • Home
    • About Us
    • Contact Us
    • Privacy Policy
    • Terms & Conditions

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