By As generative AI (GenAI) applications explode across industries, enterprise leaders are waking up to a new reality: traditional security tools weren’t designed for this. In this TFiR clip, Rupesh Chokshi, SVP and GM of Application Security at Akamai, outlined how the company is stepping up with a purpose-built “Firewall for AI” to meet today’s emerging risks.
“We’re going to live in a hybrid world… and we need protections that do both [legacy + AI workloads] simultaneously,” said Chokshi.
📹 Going on record for 2026? We're recording the TFiR Prediction Series through mid-February. If you have a bold take on where AI Infrastructure, Cloud Native, or Enterprise IT is heading—we want to hear it. [Reserve your slot
A Shift Toward AI-Native Security
According to Chokshi, the majority of Akamai’s enterprise customers are in a hybrid phase—running both traditional applications and introducing GenAI-based solutions. This hybrid reality creates new challenges for security teams that need unified, intelligent protection for vastly different workloads.
The problem isn’t just complexity—it’s also scale. New threat vectors are targeting AI-specific weaknesses like model poisoning and hallucination-based misinformation. And because many enterprise use cases involve sensitive data and critical internal workflows, the stakes are high.
“We’re seeing different threat vectors trying to go in different angles,” Chokshi noted, citing both internal and external risks.
Real-World Use Cases
Chokshi shared a compelling example from a global banking customer deploying a GenAI-powered conversational assistant for internal use. While the tool is meant to boost productivity, it still presents risks. A poisoned model, for instance, could spread misinformation internally, which could eventually reach customers.
This illustrates why enterprise security must extend beyond traditional perimeter defenses. Akamai’s Firewall for AI helps ensure safe deployment even during pre-production testing, using red teaming techniques to simulate attacks and validate resilience.
“You still run the same challenges on data poisoning… and if you start to share misinformation within the company, ultimately, you’re going to end up sharing some of that with customers,” Chokshi warned.
From Data Center to Edge
Where your AI runs matters just as much as what it does. Chokshi emphasized Akamai’s ability to support deployments anywhere—from centralized cloud to edge locations.
“We are uniquely positioned… across the continuum of compute, all the way from big data centers to the edge,” he said.
This flexibility is key in industries like healthcare or finance, where latency, compliance, and data residency drive where workloads can run. It also speaks to Akamai’s broader strategy: abstract complexity, offer deployment flexibility, and infuse AI workloads with robust, enterprise-grade protection.
Final Thoughts
Akamai isn’t just adapting existing tools for AI—it’s building new ones that recognize the very different threat model GenAI presents. The Firewall for AI reflects that strategy and is already finding traction among forward-looking enterprise clients.
For enterprise developers, DevSecOps teams, and CIOs navigating the GenAI shift, this conversation offers a real-world look into the next generation of cloud-native, AI-aware security tooling.
