Cyber Security News

Hacking Editorial Brief — September 11, 2026

AI Agents Execute Coordinated Attack on PaperCut Infrastructure

Blackpoint Cyber and GreyNoise identified an active campaign deploying hundreds of AI agents to exploit zero-day vulnerabilities in PaperCut NG and PaperCut MF print management software. The attack represents a tactical shift from individual AI-assisted exploitation to coordinated swarm tactics, with researchers documenting automated reconnaissance, vulnerability scanning, and exploitation across multiple targets simultaneously. Anthropic separately confirmed that AI capabilities now enable lone operators to sustain attack campaigns previously requiring teams of skilled personnel, validating concerns that AI tools are fundamentally altering the economics and scale of offensive operations.

Florida Driver License Database Compromised, Internet Archive Breach Exposes 31 Million Records

Threat actors claimed access to Florida's driver license database, with stolen records appearing on multiple cybercrime monitoring platforms, though the full scope and legitimacy of the breach remain under investigation. Separately, the Internet Archive confirmed a September 2026 breach exposing over 31 million files including email addresses and usernames, with pro-Palestinian actors conducting concurrent DDoS attacks against the platform. The incidents underscore continued targeting of government identity systems and large-scale data repositories.

US Lawmakers Push Sanctions Against Indian Hack-for-Hire Firms

Three U.S. lawmakers formally called on the American government to blacklist multiple Indian IT firms over documented hack-for-hire operations, escalating legislative pressure on commercial surveillance vendors. The move follows broader efforts to disrupt the mercenary spyware ecosystem and comes as the industry faces increased scrutiny over targeting of journalists, activists, and political figures. Meanwhile, commercial AI tools failed to defend Hugging Face against OpenAI's sanctioned attack during testing, with researchers concluding that open-weight models offer more effective security than proprietary detection systems in adversarial scenarios.

Sources: Hackread · SiliconANGLE · WWSB · Cybersecurity News · KFGO · Financial Times