AI Is Now a Weapon: Supply-Chain Attacks, Rogue Agents & the EU’s AI Act Crackdown
Today on Techstrong Gang, Mike Vizard hosts Sid Nag (Founder, CEO and Chief Research Officer at Tekonyx), Kate Scarcella (Cybersecurity Architect and Chair of the Continuous Delivery Foundation’s Cybersecurity Special Interest Group), and Chris Blask (Co-Founder and CEO of QuietWire). The panel breaks down three stories reshaping AI security and regulation.
AI Security Under Attack: How Fast AI Threats Move
Attackers now exploit AI systems within 48 hours of a vulnerability going public. That is the sobering finding in “AI Has Become Both Weapon and Target in Modern Cyberattacks.” Security teams must now rethink their entire patch cycle. Amazon adds another data point: a North Korea-linked group is behind multiple open source supply-chain attacks. The group exploits the trust developers place in open source packages. It slips malicious code deep into enterprise environments.
Can AI Agents Be Trusted?
The panel tackles a hard question next: can AI agents act on their own safely? The “AI Proving Grounds Consortium Tests Whether Security Agents Are Ready to Act” piece covers a new industry effort. The consortium stress-tests autonomous security agents before granting them real authority. A companion piece, “The Model Wasn’t Rogue. The Control Plane Was,” makes a sharp argument. When AI agents misbehave, a poorly governed control plane is usually the real cause, not a rogue model. This distinction matters for how enterprises deploy agentic AI.
Regulators Step In
Regulation finally catches up. The panel discusses how the EU begins enforcing its AI Act in “EU Begins Enforcing AI Act, Expanding Oversight of OpenAI, Anthropic and Google.” The move marks a major test. It shows how far governments will go to police frontier AI companies that operate across borders.
Together, these three stories show an AI ecosystem under pressure from every direction. Attackers exploit it. Builders race to give it more autonomy. Regulators move to rein it in.
