AI Security Weekly – Midweek Briefing

Palo Alto Networks acquires Protect AI, bolstering its AI-powered cybersecurity offerings.U.S. Congress passes the Take It Down Act, tackling AI-generated deepfake abuse.Texas House greenlights a new Cyber Command, aimed at defending state infrastructure.RSA Conference 2025 focuses on agentic AI and industry-wide collaboration.

Author

Andres Guillen
April 30, 2025

Palo Alto Networks Acquires Protect AI to Bolster AI Security

Palo Alto Networks has acquired Protect AI, a Seattle-based startup focused on AI pipeline security. The acquisition, reportedly valued at up to $700 million, was announced at the RSA Conference 2025. Protect AI’s platform secures the AI/ML development lifecycle against threats like data poisoning and model manipulation. Palo Alto plans to integrate these capabilities into its Prisma AIRS platform, which leverages autonomous agents for security operations.

Editor’s Commentary:
This acquisition is a strategic bet on securing the entire AI pipeline, not just runtime inference. As autonomous agents take on more critical security tasks, their integrity—and that of the models and data that drive them—becomes mission-critical. CISOs should be reevaluating their AI security stack accordingly.

U.S. Congress Passes Take It Down Act to Combat Deepfake Harms

In a rare show of bipartisan consensus, the Take It Down Act passed the House 409–2 and is now law. The act criminalizes the distribution of AI-generated non-consensual pornography and mandates social platforms to remove such content within 48 hours of notification. The law will be enforced by the FTC and has received backing from major tech companies including Meta and Snap.

Editor’s Commentary:
This is the clearest signal yet that regulators are ready to crack down on malicious AI content. While enforcement will be complex, the law sets a national precedent and forces platforms to develop AI content moderation tools that meet regulatory timelines.

Texas Approves Creation of New State Cyber Command

The Texas House approved the formation of a dedicated Cyber Command headquartered in San Antonio. With an expected budget of over $345 million over five years, the facility will include a digital forensics lab, incident response teams, and secure data infrastructure. It will be developed in collaboration with the University of Texas at San Antonio.

Editor’s Commentary:
State governments are no longer waiting for federal leadership on cybersecurity. This investment positions Texas as a model for how local entities can proactively defend against rising nation-state and ransomware threats.

RSA Conference 2025 Spotlights Agentic AI and Open Collaboration

This year’s RSA Conference highlighted how agentic AI—AI systems capable of making decisions and taking actions autonomously—is reshaping security workflows. Cisco introduced a new open-source Foundation AI Security Model, while Google detailed how threat actors are beginning to experiment with generative AI tools for reconnaissance and social engineering.

Editor’s Commentary:
AI security is now officially a two-front battle: defending AI systems while preparing for adversaries using AI offensively. Open-source models like Cisco’s are a welcome sign of collective defense, but enterprise leaders should prepare for a near-future in which AI is both attacker and defender.

Final Word

From new laws to landmark acquisitions, the midweek pulse of AI and cybersecurity shows that regulators and tech leaders alike are moving decisively. CISOs, founders, and IT leaders must track not just threats—but also the shifting terrain of accountability and automation.

Call to Action:
Subscribe to AI Security Weekly to stay ahead of the latest AI and cybersecurity breakthroughs, risks, and regulations.

Sources