Highly performing AI agents can still fail to spot deception, study finds

Large language models (LLMs), artificial intelligence systems that can process and generate texts in different languages, are now used daily by many people worldwide. As these models can rapidly source information and create convincing content for specific purposes, they are now also used in some professional settings or for gathering legal, medical, or financial information.


play_arrow All United States of America News
7 h.
Technology
ID: 949260826802966238


Similar News expand_more


1 M. Analysis: Deepfake Fraud Explodes as AI Tools Become Widely Accessible
Technology
1 M. A new method to steer AI output uncovers vulnerabilities and potential improvements
Technology
3 w. How can you avoid AI sycophancy? Keep it professional, researchers say
Science
3 w. AI is making travel scams nearly impossible to spot. Here’s what to do
Travel
1 M. 3 Questions: Using AI to accelerate the discovery and design of therapeutic drugs
Science
1 w. Researchers put six AI agents on Discord for two weeks, exposing risky failures
Science
2 w. How our AI bots are ignoring their programming and giving hackers superpowers
Technology
1 M. Indian AI lab Sarvam’s new models are a major bet on the viability of open source AI
Technology
1 M. Helping AI agents search to get the best results out of large language models
Technology
2 w. New ensemble AI model enhances cyber intrusion detection with high accuracy
Technology
3 w. OpenAI COO says ‘we have not yet really seen AI penetrate enterprise business processes’
Technology
2 w. DiligenceSquared uses AI, voice agents to make M&A research affordable
Technology
3 w. AI agents have their own social network: Moltbook study tracks topics and toxicity
Science
1 w. Can AI make you more creative? Study maps when it helps and when it slows work
Technology
2 w. What does the US military’s feud with Anthropic mean for AI used in war? | AI (artificial intelligence)
Military
1 M. I Tried RentAHuman, Where AI Agents Hired Me to Hype Their AI Startups
Technology
1 M. Indian AI lab Sarvam’s new models are a major bet on the viability of open
Technology
1 M. Q&A: AI agents created their social media platform—expert discusses what it means
Technology
4 d. AI chatbots' tendency to always agree may reinforce delusions in vulnerable users
Science
4 w. AI model edits can leak sensitive data via update 'fingerprints'
Technology
1 M. Big Tech Says Generative AI Will Save the Planet. It Doesn’t Offer Much Proof
Technology
1 M. OpenAI disbands mission alignment team, which focused on ‘safe’ and ‘trustworthy’ AI development
Technology
9 d. 'Privacy by design': Tech protects against identity leaking during AI photo editing
Technology
5 d. 'CODE RED' Explores AI's Potential to Spot the Next Minnesota
Technology
4 w. Pinpointing direction in noisy 2D data: New algorithm could improve imaging, AI, particle research and more
Technology
1 M. HHS Is Making an AI Tool to Create Hypotheses About Vaccine Injury Claims
Science
5 d. See which jobs are most threatened by AI, and who may be able to adapt
Technology
1 M. AI support spots 9% more breast cancers without raising false positives
Science
1 w. AI and work: An expert assesses how far this revolution still has to run
Technology
3 w. Researchers develop a system that detects subtle defects missed by existing industrial visual inspection
Science
3 w. Safe AI isn't enough: Fairness, honesty and transparency are needed to benefit humanity, argues researcher
Science
7 h. ‘ChatGPT gets a D’: Washington State University study finds ‘inaccuracies’ with AI
Education
1 w. Can AI read papers like a scientist? A new benchmark shows where LLMs fail
Science
1 M. Will AI put me out of a job? How artificial intelligence is being used in casino gaming
Technology
1 M. Study argues online clicks and scrolls are 'thin labor' powering AI
Technology
2 w. Luma launches creative AI agents powered by its new ‘Unified Intelligence’ models
Technology
9 d. The Illusion Of Dialogue: How Washington And Tel Aviv Turned Diplomacy Into Deception – OpEd
Politics
3 w. Perplexity’s new Computer is another bet that users need many AI models
Technology
2 w. EXCLUSIVE: Luma launches creative AI agents powered by its new ‘Unified Intelligence’ models
Technology
1 M. Interpol backroom warriors fight cyber criminals 'weaponising' AI
Crime
2 w. AI is getting smarter, but not wiser: A new roadmap aims to fix that gap
Technology
2 w. How the U.S. can maintain its edge in AI without leaving workers behind
Technology
1 M. DHS says immigration agents appear to have lied about shooting in Minnesota
Crime
2 w. Job interviews now include AI and personality tests. Here’s what candidates should know.
Science
1 M. Kana emerges from stealth with $15M to build flexible AI agents for marketers
Technology
2 w. A suite of government hacking tools targeting iPhones is now being used by cybercriminals
Technology
1 M. Xcode moves into agentic coding with deeper OpenAI and Anthropic integrations
Technology
4 d. 4 Free Agents Still Available Who Could Start for the Los Angeles Chargers
Technology
3 w. At least 6 FBI agents who worked on Trump classified documents case fired: Sources
Crime
Popular countries based on strong economic and political relations
Add Watch Country

arrow_drop_down