DataBreachToday.com

Agencies Prioritizing Tracking Use Over Enforcing Immediate Cutoffs
Federal agency staffers tell ISMG they are still using Anthropic's AI tools weeks after U.S. President Donald Trump ordered an immediate halt, as officials prioritize mapping dependencies and evaluating alternatives over enforcing a rapid shutdown.

Bridging Observability Gaps With AI, OTel and Scalable Data Models
As AI-driven development and cloud adoption accelerate system complexity, traditional observability tools are struggling to keep pace. This analysis outlines four foundational pillars to close visibility gaps and enable faster, AI-powered root cause analysis.

CISO Insights Reveal Gaps Between AI Adoption Speed and Data Security Maturity
A survey of 124 CISOs reveals most enterprises have scaled AI but lack confidence in data security controls. With only one in five initiatives meeting KPIs, gaps in enforcement, data trust and visibility are emerging as critical barriers to AI success.

Real-Time Payments, AI-Led Exploits Are Exposing Flaws Fraud Detection Can't Catch
For years, fraud prevention has followed a familiar script. A transaction is initiated. A model evaluates it. Fraud still gets detected as it happens or after it occurs. But this model is breaking down with the rise of instant payments and artificial intelligence tools.

Slumping Stock and Slower Growth Than Rival Rubrik Pave Way for Take-Private Deal
Reuters reported that Commvault is working with Goldman Sachs to explore a sale after receiving takeover interest from both private equity firms and strategic buyers. Thoma Bravo is among the buyers that have expressed interest in Commvault, sources told Reuters.

Chatbots Getting Better Making Final Diagnoses, But Clinical Reasoning Still Weak
General purpose large language model chatbots are getting better at coming up with patients' final diagnoses but are still weak in clinical reasoning, including generating differential diagnoses to identify and rule out other potential conditions and causes of symptoms.

Congressional Funding Standoff Still Unresolved
The Cybersecurity and Infrastructure Security Agency has told furloughed workers to report to work despite an ongoing funding lapse. U.S. DHS officials in recent days directed all furloughed personnel to return to work on their next scheduled shift, amid increasing concerns from cybersecurity analysts.

Malware-as-a-Service Operations Favors Russian-Speaking Customers
An emerging remote access Trojan targeting Android devices in Spanish-speaking nations is propagating fraudulent advertisements as an initial access point on Meta-owned applications.

OpenAI Unveils GPT‑5.4‑Cyber in Pointed Rejoinder to Anthropic
OpenAI unveiled Tuesday its answer to AI rival Anthropic's much-touted private release of a cybersecurity model by announcing the broader availability of GPT‑5.4‑Cyber. Internal safeguards, customer verification and "trust signals" will safeguard the world from misuse, the company asserted.

Why Cyber Risk Is Now Shaped as Much by Nations as by Hackers
Wars are becoming more frequent, and are no longer only kinetic. They are just as active in the cyber world, with impacts much larger than can be imagined. This also leads to state-sponsored hacktivists targeting the critical infrastructure of nations.

Veteran Hardware Hacker's Chip Facilitates More Trustworthy and Secure Devices
How can we trust hardware to not betray us? Enter the Baochip-1x, a piece of largely open-source silicon created by Andrew "Bunnie" Huang, which he said is designed to give developers an affordable, security-focused and attestable chip, not least for building high-assurance, embedded devices.

European Governments Grow Suspicious of Silicon Valley
French abandonment of American software for open-source alternatives continues apace, with all government ministries now facing a fall deadline for outlining plans to reduce their dependence on U.S. tech. France must "regain control of our digital destiny," said public action minister David Amiel.

Patients Allege Health Entities Did Not Get Consent to Record Conversations
Proposed federal class action litigation alleges that two California healthcare organizations violated patient privacy in their use of an AI-enabled ambient tool that records, transcribes, and processes sensitive conversations between clinicians and patients without individuals' consent.

Ex-Microsoft CIO: Mythos Could Surface Known Flaws Faster Than Vendors Can Fix Them
Former Microsoft CIO Jim DuBois and IDC's Frank Dickson say Claude Mythos Preview could rapidly surface long-known but unfixed software flaws at scale, forcing vendors and enterprises to strengthen patch validation, orchestration and deployment before attackers exploit the backlog.

Operational Discipline and Judgment Are Critical in Managing Cyber Risk
Transitioning from armed forces can feel like stepping into unfamiliar terrain. Nowhere is this perception stronger than in cybersecurity. The good news: Many of the skills veterans have already developed translate directly to cybersecurity roles.

Anthropic's AI Model Exposes How Unprepared Enterprises Are to Respond
Anthropic's announcement this week of Claude Mythos Preview frontier model capable of finding zero-days flaws humans may miss is both a warning and a call to action for CIOs: The way enterprises have been managing cybersecurity is about to change forever, and they need to get ready.

Early Tests of New Anthropic AI Model Show Fast Detection, Better Flaw Correlation
CrowdStrike's early testing of Anthropic's new Claude Mythos Preview AI model shows faster vulnerability detection and improved cross-system context, signaling a shift toward AI-driven security operations that compress discovery-to-response timelines and force new defensive frameworks.