Toggle light / dark theme

Microsoft finds security flaw in AI chatbots that could expose conversation topics

Your conversations with AI assistants such as ChatGPT and Google Gemini may not be as private as you think they are. Microsoft has revealed a serious flaw in the large language models (LLMs) that power these AI services, potentially exposing the topic of your conversations with them. Researchers dubbed the vulnerability “Whisper Leak” and found it affects nearly all the models they tested.

When you chat with AI assistants built into major search engines or apps, the information is protected by TLS (Transport Layer Security), the same used for online banking. These secure connections stop would-be eavesdroppers from reading the words you type. However, Microsoft discovered that the metadata (how your messages are traveling across the internet) remains visible. Whisper Leak doesn’t break encryption, but it takes advantage of what encryption cannot hide.

DBS Rolls Out GenAI Chatbot, Southeast Asia’s Largest Bank Incorporates AI In Its Workflow

In today’s AI news, Southeast Asia’s largest bank is rolling out an AI chatbot for its corporate clients, giving “round-the-block” access for customer care needs. A pilot version of the generative AI-powered chatbot, named DBS Joy, was rolled out in February. It has since managed over 120,000 unique chats, DBS claimed in a statement. The virtual assistant also cut waiting times, and customer satisfaction scores rose by 23%. The bank’s virtual assistant was first rolled out in 2018, under the same name.

In other advancements, Artificial intelligence giant OpenAI on Monday announced a new initiative that aims to make it easier for service members and veterans to use AI tools when they’re transitioning from military service to the workforce. The ChatGPT-maker announced that service members within 12 months of separation or retirement from, or any veteran within their first year of leaving military service, can access a free year of access to ChatGPT Plus, the company’s subscription-based tool.

Meanwhile, Neurodiverse professionals may see unique benefits from artificial intelligence tools and agents, research suggests. With AI agent creation booming in 2025, people with conditions like ADHD, autism, dyslexia and more report a more level playing field in the workplace thanks to generative AI. A recent study from the UK’s Department for Business and Trade found that neurodiverse workers were 25% more satisfied with AI assistants and were more likely to recommend the tool than neurotypical respondents.

And, San Francisco became a meme — a symbol of American urban decay. Between 2019 and 2024, the city lost 4 per cent of its population, one of the largest declines in America. But thanks to AI — and new “tough on crime” mayor Daniel Lurie — the vibe has, unquestionably, shifted. Since 2020, more than 2,400 AI companies have been founded in San Francisco, a city of just 830,000 people. “Hacker houses”, where bright-eyed coders live communally and build what they hope will be the next Google, have cropped up across the city.

In videos, Six of the most influential minds in artificial intelligence joined FT Live for an exclusive conversation on how their breakthroughs and the current state of AI are shaping our world. Jensen Huang, Yoshua Bengio, Geoffrey Hinton, Fei-Fei Li, Yann LeCun, and Bill Dally spoke with the FT’s AI editor, Madhumita Murgia at the FT Future of AI Summit in London. Together, they reflected on decades of pioneering work — from neural networks to generative AI and discuss the ethical, social, and economic implications of the technology they helped to create.

Then, Ellis Hamburger and Alex Heath chat about the Sources launch party, and living in late-stage extractionism. Then, they are then joined by Runway CEO Cristobal Valenzuela to discuss the hardest part about AI, why golf courses are a bigger problem than data centers, and how world models are changing our world forever.

And, David Sacks, White House AI and Crypto Czar, joins Marc, Ben, and Erik on the a16z podcast to explore what’s really happening inside the Trump administration’s AI and crypto strategy. They expose the regulatory capture playbook being pushed by certain AI companies, explain why open source is America’s secret weapon, and detail the infrastructure crisis that could determine who wins the global AI race.

Apache OpenOffice disputes data breach claims by ransomware gang

The Apache Software Foundation disputes claims that its OpenOffice project suffered an Akira ransomware attack, after the threat actors claimed to have stolen 23 GB of corporate documents.

Apache OpenOffice is a free, open-source office suite that includes word processing, spreadsheets, presentations, graphics, and database tools. It’s compatible with major file formats, such as Word and Excel, and runs on multiple operating systems.

On October 30th, the Akira ransomware gang claimed it had breached Apache OpenOffice and stolen 23 GB of data, including employee and financial information, as well as internal files.

Malicious Android apps on Google Play downloaded 42 million times

Hundreds of malicious Android apps on Google Play were downloaded more than 40 million times between June 2024 and May 2025, notes a report from cloud security company Zscaler.

During the same period, the company observed a 67% year-over-year growth in malware targeting mobile devices, with spyware and banking trojans being a prevalent risk.

Telemetry data shows that threat actors are shifting from traditional card fraud to exploiting mobile payments using phishing, smishing, SIM-swapping, and payment scams.

Brains and stock markets follow the same rules in crisis, study finds

What do brains and the stock market have in common? While this might sound like a set-up for a joke, new research from U-M researchers reveals that the behaviors of brains and economies during crises can be explained using observations common in the realm of physics. Their work is published in the journal Proceedings of the National Academy of Sciences.

UnCheol Lee, Ph.D. of the U-M Department of Anesthesiology and his collaborative team came up with the idea upon observing that some patients under anesthesia recover faster than others.

“Anesthetic drugs can be considered as introducing a controlled crisis in the brain, interrupting the brain’s network to induce unconsciousness,” explained Lee.

“Only a Question of Time” Does AI Mean We’re DOOMED? Plus Oz Pearlman Reads Piers Morgan’s Mind!

ExpressVPN: Right now you can get an extra four months of ExpressVPN for free. Just scan the QR code on the screen, or go to https://ExpressVPN.com/PIERS and get four extra months for free.

Two years ago, Elon Musk was among a thousand experts to sign an open letter demanding an urgent pause on the advancement of Artificial Intelligence because of the risks concerning job losses, misinformation and more.

But now Musk is now spending a billion dollars a month to compete in an AI arms race, which is inflating the stock market to bursting point.

Amazon just laid off 14,000 workers in its ongoing A.I pivot — so, are the worst fears of doomsaying experts already coming true?

Joining Piers Morgan to discuss are respected thinkers in this field; Dr Roman Yampolskiy, Dr Michio Kaku, Alex Smola and Avi Loeb.

Then; he’s performed for presidents, billionaires and sports stars, but Oz Pearlman’s recent extraction of Joe Rogan’s pin number may have been his biggest hit yet.

Massive surge of NFC relay malware steals Europeans’ credit cards

Near-Field Communication (NFC) relay malware has grown massively popular in Eastern Europe, with researchers discovering over 760 malicious Android apps using the technique to steal people’s payment card information in the past few months.

Contrary to the traditional banking trojans that use overlays to steal banking credentials or remote access tools to perform fraudulent transactions, NFC malware abuses Android’s Host Card Emulation (HCE) to emulate or steal contactless credit card and payment data.

They capture EMV fields, respond to APDU commands from a POS terminal with attacker-controlled replies, or forward terminal requests to a remote server, which crafts the proper APDU responses to enable payments at the terminal without the physical cardholder present.

This Chip Computes With Light, Breaking the 10 GHz Barrier for AI

Researchers have developed an optical computing system that performs feature extraction for quantitative trading with unprecedentedly low latency. Many advanced artificial intelligence (AI) systems, including those used in surgical robotics and high-speed financial trading, rely on processing lar

/* */