1. LIVE: Acting Attorney General Blanche testifies before Senate — Tuesday 19 May 2026
2. Acting U.S. Attorney General Todd Blanche appeared before a Senate subcommittee to outline the Justice Department's proposed budget for fiscal year 2027. The live testimony focused on funding priorities that will shape federal law enforcement, prosecution efforts, and legal operations in the years ahead. Subcommittee members questioned Blanche on how the department plans to allocate resources amid shifting national security demands and evolving criminal threats.
   
   The proposed budget covers critical areas such as counterterrorism, cybercrime investigations, and civil rights enforcement. Blanche emphasized the need for sustained investment in technology and personnel to keep pace with increasingly sophisticated criminal networks. Senators from both parties raised concerns about oversight mechanisms and the department's ability to maintain independence while addressing high-profile cases that often draw international attention.
   
   For global audiences, the outcome of these deliberations carries weight beyond U.S. borders. Adequate funding for international law enforcement cooperation could strengthen efforts against cross-border trafficking, financial crimes, and foreign influence operations. Observers in Europe and Asia are watching closely, as U.S. Justice Department capacity directly affects joint investigations and extradition processes that rely on American resources and legal frameworks.
3. Watch the full video from Reuters below.

1. Everyone's Getting Hacked" – Matthew Berman on the Rising AI-Driven Cyber Threat Landscape (May 2026)
2. In his recent video, AI commentator Matthew Berman warns that cybercrime is becoming easier, more scalable, and more profitable than ever—largely due to advances in artificial intelligence. While AI offers powerful defensive tools, malicious actors are already leveraging it for attacks, and the trend is accelerating.Key Recent Incidents Highlighting the Problem
3. Google's Discovery of AI-Generated Zero-Days: Google's Threat Intelligence Group (GTIG) reported the first known instance of a threat actor using an AI-developed zero-day exploit in the wild. Zero-days (previously unknown vulnerabilities) are highly valuable because they remain undetected until used. AI accelerates their discovery by systematically analyzing code—especially open-source code—at scale. Google’s proactive countermeasures reportedly prevented a larger attack.
4. Shy Halud npm Worm: A sophisticated supply-chain attack involving a "dead man's switch" that wipes directories upon token revocation. It has spread across hundreds of npm packages (including major ones) and into PyPI, replicating like a worm. This is linked to broader credential theft from prior attacks (e.g., Team PCP) and amplified by AI tools that help generate and deploy such malware.
5. Vercel Breach: The company explicitly stated that AI almost certainly accelerated the attack. An employee was compromised via a third-party AI platform, allowing sophisticated intruders to move quickly with deep knowledge of Vercel’s systems.
6. How AI is Supercharging AttacksBerman outlines several ways AI empowers adversaries:
7. Vulnerability Discovery & Exploit Generation: AI runs 24/7, understands code deeply, and finds bugs humans miss. It’s not creating vulnerabilities (those come from human error) but rapidly uncovering them.
8. Malware Development & Evasion: AI speeds up creation of polymorphic malware, obfuscation networks, and decoy logic. Autonomous agent setups allow malware to operate indefinitely with near-zero marginal cost (especially open-source/local models).
9. Supply Chain & Initial Access: Targeting AI coding environments and dependencies for broader network pivots, ransomware, etc.
10. Phishing & Social Engineering: Deepfakes and sophisticated phishing are rising. Berman recommends family code words/phrases as a low-tech defense against voice/video impersonation.
11. Malicious actors often avoid the latest frontier models (like GPT-5.5 Cyber or unreleased ones) due to strong guardrails. Instead, they use smaller, less-restricted models, distillation techniques, or anonymized access to scale operations.The Defense Side: "Your AI vs. My AI"A core theme is an arms race where superior AI wins:
12. Frontier Labs’ Responses:
13. Anthropic’s Mythos (Project Glasswing, ~10T parameters): A powerful cyber model not publicly released due to risks. Shared with select partners (AWS, Google, Apple, etc.) to harden systems first. It uncovered old zero-days in OpenBSD, FFmpeg, and Linux kernel chains.
14. OpenAI’s GPT-5.5 Cyber & Daybreak: More open approach—trusted access for defenders, iterative releases, and tools to find/fix vulnerabilities faster.
15. Google and others are advancing similar defensive AI.
16. 
    
17. Economic & Practical Advantages for Defenders: Building world-class models requires massive compute, data, energy, and talent—feasible for well-resourced companies and nations but not most rogue groups. Better models can find and patch vulnerabilities that weaker attacker models miss. Jensen Huang (NVIDIA) echoed this "my AI vs. your AI" dynamic.
18. State actors (e.g., China, North Korea) remain a wildcard with resources to compete at the top level.Berman’s Takeaways and AdviceThe situation is concerning but not hopeless. AI amplifies both offense and defense, but economics and scale favor organized good actors (companies/governments) over most criminals. Key recommendations include:
19. Use secure AI agent setups (e.g., isolated environments like GenSpark Claw, which Berman promotes).
20. Review dependencies and avoid unchecked "vibe coding."
21. Stay vigilant on phishing/deepfakes.
22. Support proactive security from frontier labs.

Overall, Berman argues we’re in an early phase of AI-amplified cyber threats. The volume, velocity, and sophistication will increase, but defensive AI—properly deployed—can outpace attackers. He urges viewers to like, subscribe, and explore tools for safer AI use.This video serves as a timely wake-up call amid rapid AI progress in 2026, blending alarming examples with reasoned optimism about technological countermeasures. For full details, watch the original on Matthew Berman’s channel.

https://www.youtube.com/watch?v=hAzhVloGkOw