A TechRadar article says deploying AI agents requires governed enterprise data first, including classified inputs, least-privilege access, provenance and retained records of prompts, outputs and decisions.
ESET's 2026 SMB Cyber Risk Report found 49% of surveyed UK small businesses suffered a cyber incident in the past year, with AI-powered malware their top concern and phishing, unpatched flaws, weak passwords and poor monitoring the leading causes.
President Trump and President Xi Jinping meet in Washington with AI safety and incident communication on the agenda, while U.S. debate over an AI freeze and Senate restrictions on advanced tools show the difficulty of regulating fast-moving AI.
Domain registrar Identity Digital has spun out Known Systems AI, an independent company that will push the DNSid framework as an open standard tying AI agents to the organizations that create them.
Cisco Talos released an open-source toolkit for hunting AI-enabled malware and detailed CLOSEDQUORUM, a Windows credential stealer with no command-and-control server that puts each tactical decision to a majority vote of four commercial language models.
Google's Find Hub tracking network can locate a missing Android phone, tablet or accessory, and does not require a second Android device. It also allows remote locking, sound playback and, as a last resort, a remote data wipe.
NPR News reports that a Southeast Asia cybercrime operation trapped computer engineers into defrauding victims of billions of dollars, with the scammers themselves described as victims of a "modern slavery situation." Wired's Andy Greenberg explains the operation and a scam involving Fresh Air.
A report by Refuge and End Violence Against Women, reported by The Guardian, finds online abuse and misogyny have become an accepted risk for women, with police and platforms often failing to act. A survey of 2,000 women found 17% of police complaints were not investigated.
TechRadar has named NymVPN its VPN Deal of the Week after the provider introduced lower-commitment price plans, which the publication describes as a cheaper way to try its mixnet security.
Policy & Regulation · Cybersecurity · Large Language Models
OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei are expected to address the U.N. Security Council on Wednesday in New York, a meeting on artificial intelligence that comes as safety concerns mount and President Trump vows to encourage the technology.
New SpyCloud research found password-stealing malware has exposed credentials from 1,787 U.S. water and wastewater organizations, with at least 250 appearing to allow access to operational networks, as the sector faces separate Iran-linked hacks.
A developer released Drop, a language-independent Linux sandbox that runs third-party programs without root by using namespaces and an optional gVisor user-space kernel.