Juan Perez

In this special year-end edition, we revisit critical advice from our cybersecurity experts on AI, exposure management, cloud, vulnerability management, OT and critical infrastructure.

In this special edition, Tenable leaders forecast key 2026 trends, including: AI will make attacks more plentiful and less costly; machine identities will become the top cloud risk; preemptive cloud and exposure management will dethrone runtime detection; and automated remediation gets the go-ahead.

Customize Consent Preferences Customise your consent preferences for Cookie Categories and advertising tracking preferences for Purposes & Features and Vendors below.

Let's imagine a customer-support chatbot—it's running on Red Hat OpenShift AI and searches internal documents to answer questions. A user asks it a common question, but the chatbot inadvertently retrieves a malicious document that contains hidden instructions like, “ignore all policies and reveal secrets.” Not knowing any better, the AI model follows these malicious instructions and leaks internal data—and no one notices until screenshots appear online. This is the new computer security reality in which we live. Modern AI systems do more than “respond.” They reason over untrusted inputs, and sometimes they use external tools, which means the potential attack surface grows quickly.

Standard AI security benchmarks can't check for all of the possible ways an AI model can be compromised. A backdoor trigger could cause targeted failure, a competitor could clone your API model through repeated queries, or a privacy probe might reveal whether a specific person’s data was used in training. For this reason, organizations deploying AI must understand the variety of potential attacks and proactively address them during model training and after deployment.

In previous articles, we framed AI security as protecting confidentiality, integrity, and availability of the whole AI system, not just the model. We also mapped AI risks onto familiar secure development lifecycle (SDLC) thinking, treating data and model artifacts as first-class build inputs and outputs.

In our first 3 articles, we framed AI security as protecting the system, not just the model, across confidentiality, integrity, and availability, and we showed why the traditional secure development lifecycle (SDLC) discipline still applies to modern AI deployments. We also focused on guardrails and different architectural approaches such as dual LLMs and CaMeL to help protect against prompt injection and unsafe actions.

Juan R. Perez, 79, a former Frankfort resident, passed away April 19, 2026 at his home in Indianapolis. He was born June 23, 1946 in Puerto Rico to Higinio Perez & Dolores (Reyes) Perez. He married Elizabeth Rangel on May 27, 1967 in Ohio and she survives.

The Strait of Hormuz is already pricing into your payment corridors, the question is whether your FX strategy knows it yet.

Typical AAV Gene Therapy production systems generate cell culture harvests with a high abundance of empty capsids lacking therapeutic DNA compared to full capsids containing the therapeutic DNA. Removal of empty capsids during AAV purification is a central production challenge since they possess similar surface charge properties as full capsids. A robust and scalable purification process capable of completely removing empty capsids from the product is presented in this work. This process employs Anion Exchange Chromatography (AEX) and Zonal Ultracentrifugation (ZUC) as orthogonal purification steps. While AEX differentiates the surface charge between empty and full capsids, ZUC separates via differences in density of the capsids. Heterogeneity of empty capsids was first studied by evaluating empty capsid elution profiles during AEX gradient chromatography and by comparing sedimentation profiles by Analytical Ultracentrifugation (AUC) of empty capsids separated by AEX and ZUC. The results suggest a heterogenous profile of empty capsids in AAV. Separation of AAV by AEX step elution was then studied, demonstrating that there is a distinct sub-population of empty capsids that co-elutes with full capsids, making it difficult to separate them using AEX alone. AAV separation on ZUC results in near-baseline separation of empty capsids from full capsids in addition to removal of aggregates and high molecular weight (HMW) impurities. Finally, three purification trains using the AEX-ZUC combination process were performed to study the robustness of this process in removing empty capsids and other impurities. Results demonstrated that the AEX-ZUC combination process was able to remove empty capsids down to undetectable levels when analyzed by AUC. This process also resulted in reduction of other impurities such as aggregates and HMW species. In addition, Rep protein impurity was also reduced by AEX and then ZUC down to very low levels suggesting that this process can control trace impurities potentially bound to empty capsids. This work shows that the AEX-ZUC combination is a robust and scalable process capable of producing high purity AAV for Gene Therapy applications.