Industry Insider

AI Agent Security Gaps Widen: Report

Organisations are deploying AI agents without adequate controls to govern them, according to new research from Rubrik Zero Labs.

Frozen in COVID: Victoria Police's 9-month FOI delays

Victoria Police is taking an average of nine months to process Freedom of Information requests – unchanged since at least 2022, when it was attributed to COVID-19.

Guide to PDF ingestion for AI

A detailed guide clarifying how AI systems should process PDF documents has been published by The PDF Association. The FAQ warns that common ingestion practices risk information loss and hallucinations.

NSW Treasury insider charged over 5,600 sensitive files

A NSW Treasury employee has been arrested and charged over the alleged theft of more than 5,600 sensitive government documents in what the state government has declared a significant cyber incident.

Poor AI Outcomes Tied to Weak Data Governance, Gartner

Organisations that achieve positive AI outcomes invest up to four times more in data quality, governance, and workforce preparation than those with poor AI results, according to new research from Gartner.

Singapore pushes global GenAI testing benchmark

Singapore has put forward a proposed international standard, ISO/IEC 42119-8, that would set standardised testing approaches for generative AI systems, including benchmarking and red teaming methodologies.

Claude Mythos and Project Glasswing: why an AI superhacker has the tech world on alert

New, more powerful artificial intelligence (AI) models are announced pretty regularly these days: the latest version of ChatGPT or Claude or Gemini always has new features and new capabilities that its makers are eager for customers to try out. But now Anthropic has announced a new model with great fanfare, but is only giving access to a select handful of users. In what the New York Times calls a “terrifying warning sign” of the model’s power, the company has instead started an initiative called Project Glasswing to use the model for good instead of evil.

[more featured articles]

Breaking news

Agentic threats prompt IBM enterprise defence rethink

IBM has launched two new cybersecurity offerings aimed at helping organisations defend against a rising class of AI-generated threats, as attackers begin using frontier AI models to automate and accelerate attacks.

Alliance Targets Data Access Complexity

Two enterprise data management companies have announced a partnership to simplify data access and cut costs for AI initiatives. Solix Technologies will integrate VirtualZ Computing’s no-code data access technology into its Common Data Platform (CDP).

OpenAI workspace agents target enterprise AI race

OpenAI has introduced workspace agents in ChatGPT, enabling business teams to build shared AI agents that automate complex, long-running workflows across an organisation.

Security Firm Targets Credential Theft

Detecting credential theft as it happens is the aim of Decipio, a new community security tool from Arctic Wolf. The tool targets a well-documented Windows attack vector, aiming to alert defenders before stolen credentials are weaponised inside a network.

Cloud File Vendor Nasuni Retools for the AI Era

Governing AI agent access to corporate file data has emerged as a pressing challenge for enterprise security and compliance teams. Boston-based Nasuni is attempting to address it with a new platform capability that lets AI tools operate directly on file data within existing permission structures - no separate data pipelines required.

[more breaking news]