Frontier AI on the Defender’s Side of the Application

Cybersecurity has always been an asymmetric fight. Attackers only need one successful entry point, while defenders must secure everything. AI is accelerating that imbalance. The same frontier models that can help identify vulnerabilities, analyze malware, reason through attack paths, and generate exploit strategies are now available to both attackers and defenders. The reality is that attackers are already experimenting with these capabilities at scale.

The question is no longer whether AI will change cybersecurity. It already has. The more important question is whether defenders can move fast enough to keep up.

That is why I’m excited to share that A10 Networks has been granted access to OpenAI’s Trusted Access for Cyber (TAC) program as part of a pilot initiative designed to place frontier AI capabilities into the hands of trusted cybersecurity organizations. Through TAC, A10 will use GPT-5.5 to advance defensive security research across several areas, including:

• AI red teaming and adversarial testing
• Malware analysis and threat research
• DDoS and threat intelligence investigations
• Application, API, and AI security hardening
• Faster defensive analysis and incident response workflows

Just as importantly, this is a collaborative effort. As part of the TAC program, we will regularly share findings, operational lessons, attack patterns, and safety insights with OpenAI to help strengthen the broader security ecosystem around frontier AI.

Why this Matters for A10 and the Broader Ecosystem

This collaboration is a natural extension of where A10 has already been investing. Modern enterprises are operating across hybrid environments that span applications, APIs, cloud infrastructure, and increasingly, AI systems themselves. At the same time, attackers are using automation and AI to compress the time between vulnerability discovery and exploitation from weeks into minutes. Traditional security approaches alone cannot keep pace with that kind of velocity.

A10’s security portfolio, including WAAP, DDoS defense, AI firewall capabilities, and threat intelligence is built around reducing that exposure and helping organizations defend modern application environments wherever they run. But reducing exposure is only part of the challenge.

Security teams today are overwhelmed by the scale of telemetry, alerts, malware variants, anomalous behaviors, and constantly evolving attack techniques. One of the most important opportunities with frontier AI is the ability to help defenders process and investigate that information faster, while still keeping human expertise at the center of decision-making. We see immediate opportunities in several areas.

First, building safer software internally. GPT-5.5 can assist engineering and security teams throughout the development lifecycle — from architecture reviews and secure design validation to code review, vulnerability discovery, and adversarial testing before products ever reach customer environments.

Second, advancing malware and threat intelligence research. AI-assisted analysis can help accelerate understanding of attacker infrastructure, malware evolution, evasion techniques, and emerging campaigns. Combined with A10’s expertise in DDoS and network intelligence, this improves our ability to identify and respond to threats more quickly.

Third, strengthening protections for customer environments. Effective WAAP and AI security require understanding how attackers behave. GPT-5.5 enables more realistic simulation of attack chains, prompt injection attempts, evasive behavior, and adversarial workflows that can be used to pressure-test and improve defensive systems.

And finally, accelerating investigations without removing the human from the loop. AI can dramatically reduce the manual burden involved in triage, enrichment, summarization, and contextual analysis, allowing security teams to focus more attention on higher-confidence decisions and response actions.

Defenders Win as a Community

What makes programs like TAC especially important is the recognition that cybersecurity is fundamentally collaborative. Attackers already share tooling, techniques, infrastructure, and operational knowledge. Defenders need the ability to collaborate with equal speed and sophistication. Sharing research and operational learnings responsibly across trusted organizations ultimately strengthens the resilience of the broader ecosystem.

The industry has gone through major platform shifts before — from on-premises to cloud, from perimeter security to zero trust, from static signatures to behavioral detection. AI represents another one of those shifts, except this one is moving significantly faster. Attackers already have access to frontier AI capabilities. Through OpenAI’s Trusted Access for Cyber program, defenders do too.

A10 intends to use GPT-5.5 the same way we approach every security challenge: responsibly, collaboratively, and with a focus on helping customers secure the applications, APIs, networks, and AI systems that modern businesses increasingly depend on.