Researchers get AI'drunk' to expose new cyber security risks in chatbots Fine-tuning large language models using drunken text makes them more likely to leak secrets and answer harmful questions, new UNSW-led research shows. UNSW researchers have shown that large language models (LLMs) can be pushed to imitate drunken behaviour - and that once they do, they are significantly more likely to leak confidential information and answer questions they are designed to refuse. In the study, conceptualised and led by Dr Aditya Joshi, Senior Lecturer, School of Computer Science and Engineering, with Anudeex Shetty (Research Assistant) and Professor Salil Kanhere, researchers from UNSW's School of Computer Science and Engineering (CSE) found that AI trained or prompted to mimic drunken speech patterns became more vulnerable to jailbreaking and privacy breaches than their'sober' counterparts. The UNSW team tested three methods of inducing'drunk' behaviour in LLMs. The vulnerability wasn't tested across every LLM on the market.

I ranked six AI agents--from Meta's Muse to Anthropic's Claude Duet--on how well they completed basic tasks. Last year, I remember sitting in WIRED's office, complaining to my colleagues nonstop about how much AI agents suck. They were just bad at everything. As a software reporter, I sat through meeting after meeting where tech bros --mostly men--demonstrated how their new agent would revolutionize my everyday life and automate away the tedium. The catch was that none of these so-called "agents" actually worked. Fast-forward to today, and always-on AI agents are booming, with major players racing to release their own versions to attract early adopters.

OpenAI has announced that its AI models have made several scientific discoveries in mathematics, computer science and physics. The AI giant said on Monday that the three new discoveries concerned the fields of harmonic analysis, complexity theory and quantum field theory. Researchers in harmonic analysis, a branch of mathematics that studies how complex signals and functions can be broken down into elementary waves, had previously only proved the theory in three dimensions. Alex Iosevich, a professor of mathematics at the University of Rochester who specialises in harmonic analysis, said OpenAI appeared to have "cracked" the conjecture in a "historic development". "The conjecture was formulated by my PhD adviser, Chris Sogge, so it is near and dear to my heart, and the news was a shock at first," Iosevich told Al Jazeera.

This material may not be published, broadcast, rewritten, or redistributed. Quotes displayed in real-time or delayed by at least 15 minutes. Market data provided by Factset . Powered and implemented by FactSet Digital Solutions . Mutual Fund and ETF data provided by LSEG . Why some people are falling for Meta's Muse AI Google's new AI shopping cart tracks deals across stores Your phone can alert you when prices drop; here's how Hackers could already be inside American utilities, cyber experts warn of'Trojan Horse' threat'Gutfeld!': AI offers different answers to race preference question FBI is taking an'offensive and defensive approach' to cybercrime: Director Kash Patel'The Five': Would you read your partner's ChatGPT messages?

Safety Net Look Up Versus Trending Now Say More Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Creator Playbook Mashable Selects Getting Technical with Mashable Gift Ideas For Everyone On Your List In My Bag Mashable @Home All Series CapCut is adding a ChatGPT plugin. Creators will be able to start a video edit in ChatGPT and keep refining it by chatting. Chance Townsend is the General Assignments Editor at Mashable, covering tech, video games, dating apps, digital culture, and whatever else comes his way. He has a Master's in Journalism from the University of North Texas and is a proud orange cat father. His writing has also appeared in PC Mag and .

Plus: Ukraine has regained high-speed Starlink and Starshield connectivity. We have historically allowed debates over technology to be defined by the worst thing a product might do, with remarkably little discussion of the insidious damage it does every day. A third of teens now choose to discuss serious matters with AI companions instead of other people. Students are relying on chatbots to write their papers, while social media feeds engineered to provoke envy keep users watching. These products can meet narrow safety requirements yet still erode how we think, learn, and interact with one another. As we enter an era defined by powerful AI products, we should demand that the tech industry do more than address only catastrophic harms.

AI is creating jobs as well as erasing them - but how rewarding are they? Don't believe the hype about AI taking over all forms of work . Most of the evidence to date suggests large-scale job losses are limited . But what is changing is the way work is organised. One area of increased employment is in people "cleaning up" after large language models (LLMs), or quality controlling them.

Anthropic said its AI model found a loophole in an old version of the Georgia state agency's website. Anthropic said its AI model found a loophole in an old version of the Georgia state agency's website. Mon 26 Oct 2026 07.03 EDTLast modified on Mon 26 Oct 2026 07.04 EDT Anthropic's Claude model accessed voter data in the US state of Georgia, the AI company has confirmed, in the latest example of unauthorised behaviour by the technology. The AI model managed to bypass a paywall to reach data that is publicly available but is typically accessible only after paying a fee. Claude accessed aggregate vote counts as part of a statistics project at Anthropic, which is preparing for a $1.5tn (£1.1tn) flotation.

Follow this section to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW? Smart Alerts: Get notified about major news as it happens. Follow this tag to personalize your feed and get instant alerts. Follow Go to your personalized feed WHY FOLLOW?

The Dell Laptop Buying Guide UK: Which Dell Laptop Should You Buy? When you purchase through links in our articles, we may earn a small commission. The MSI Pro Max Edge AI+ stands out with its minimalist 4L aluminum chassis and proprietary Frozr AI Pro cooling system. The MSI Pro Max Edge AI+ is a powerful compact workstation for local AI and large-scale media projects. The integrated power supply keeps the setup tidy. In the benchmark test carried out, the cooling system operated without thermal throttling. Weaknesses include the soldered-in RAM and the lack of OCuLink. The high price is an even greater drawback.