A new Cursor study reports that newer coding agents often retrieve known fixes instead of deriving them, inflating popular benchmark scores. Reward hacking means a model earns the reward without doing the intended work. Here the reward is a passing test. The intended work is deriving the bug fix. The research study focuses on agentic...Read More
The US government has allowed Anthropic to release its powerful Mythos AI model to select companies and organizations, revising license requirements after ordering an export block earlier this month in the wake of national security fears. Since the export ban earlier in June, “Anthropic has worked with the US government to address risks associated with...Read More
501 Benjamin Joseph Bertrand, an 18-year-old resident of Franklin, Louisiana, was arrested on June 25, 2026, at 12:24 PM on a warrant for distribution of child sexual abuse material. St. Mary Parish Sheriff Gary Driskell reported that deputies executed a search warrant that led to Bertrand’s arrest on cybercrime charges. The investigation began on June...Read More
Artificial Intelligence & Machine Learning , Fraud Management & Cybercrime , Governance & Risk Management Also: AI Model for Drug Development Allegedly Stolen; Accenture’s Dragos Deal Mathew J. Schwartz (euroinfosec) • June 26, 2026 Clockwise, from top left: Mathew Schwartz, Michael Novinson, Marianne Kolbasuk McGee and Chris Riotta In this week’s panel, four...Read More
ChatGPT maker OpenAI said Friday it is restricting the release of its new artificial intelligence model at the request of President Donald Trump’s administration, the latest in an unprecedented government vetting of AI products for cybersecurity risks. OpenAI said its new AI product, called GPT-5.6 Sol, would only be available for now to a “small...Read More
Germany has emerged as the European country most heavily affected by ransomware attacks, according to cybersecurity expert Christian Dörr of the Hasso Plattner Institute (HPI) in Potsdam. The alarming trend highlights the increasing threat that cybercriminals pose to businesses, government institutions, and critical infrastructure across the region. Mr. Dörr made these observations ahead of the...Read More
As detailed in Bleeping Computer, Polish authorities have arrested four individuals suspected of being part of an organized cybercrime group responsible for SIM-swapping attacks and significant cryptocurrency theft.The operation, a collaboration between Poland’s Cybercrime Bureau (CBZC) and U.S. agencies including the FBI and Homeland Security Investigations, targeted a group accused of breaching telecommunications partners and...Read More
Executive Summary The emergence of the Mistic backdoor, attributed to the notorious ransomware access broker KongTuke (also known as Woodgnat), marks a significant escalation in the sophistication and stealth of initial access operations targeting enterprise environments. Mistic is engineered for covert persistence, leveraging advanced in-memory execution, modular expansion, and anti-forensic techniques to evade detection and facilitate the sale...Read More
Italian rail company Trenitalia is informing some customers via email that it has “detected a cybersecurity incident caused by unidentified external parties” that resulted in “unauthorized access to some personal data related to travel tickets.” The company states that, following checks, it was able to “identify the affected customers” and send them the notification. ...Read More
A new Chinese artificial intelligence model is raising cybersecurity fears over its advanced capabilities, which match recent versions of giants like OpenAI and Anthropic. Unsplash A new Chinese artificial intelligence model is raising cybersecurity fears due to its advanced capabilities, which appear to be on par with Claude Opus 4.8 and OpenAI’s GPT-5.5, Axios reported....Read More
