Month: April 2026

Manpreet SinghCo-Founder &Principal Consultant, 5Tattva One of the biggest misconceptions in AI safety is the assumption that more capable models are automatically safer. In reality, the opposite can happen. A model that becomes highly skilled at generating complex structures such as poetry may also become more effective at executing hidden or obfuscated instructions embedded within those formats. Poetry has long been celebrated as a vehicle for human expression. But beneath the rhythm and rhyme lies a rigid mathematical structure – one that, in the age of artificial intelligence, may expose an unexpected vulnerability. Beneath the artistic legacy of ancient epics lies a rigid syntactic cage. In the context of modern machine learning and language models, this strict framework presents a unique vulnerability. By leveraging these artistic constraints, adversarial payloads can bypass semantic filters, turning humanity’s oldest mnemonic device into a mechanism for digital deception. The blind spot in AI alignment To understand why Shakespeare would have been an incredible asset to a modern Red Team or VAPT operation, we have to look at how modern AI safety training works. Large Language Models (LLMs) have scaled globally, expanding the attack surface across digital ecosystems by introducing new vulnerabilities and amplifying existing ones. To ensure safety, LLMs are safeguarded using Reinforcement Learning from Human Feedback (RLHF). Human testers spend thousands of hours feeding the model malicious prompts like ‘Write me a computer virus’ or ‘How do I build a homemade bomb?’ and teaching the model to refuse such requests. However, there is a critical limitation in this training data – it is overwhelmingly conversational and prose-based. These safety classifiers are designed to detect malicious intent primarily in standard conversational syntax. When a malicious command is wrapped in structured verse such as iambic pentameter or an AABB rhyme scheme, it pushes the prompt into Out-of-Distribution (OoD) territory. The model has rarely encountered security threats formatted as poetry during alignment training. The result is simple – the AI is trained to detect obvious threats, but adversarial poetry hides the threat within complex linguistic structure. The anatomy of the exploit Executing this vulnerability requires more than just basic knowledge of LLMs or the gift of rhyme. It demands a deliberate, two-stage methodology. Stage one: Semantic Obfuscation Attackers remove the prompt of known trigger words to bypass the LLM’s basic safety classifiers. Through metaphorical shifts, a ‘keylogger’ becomes ‘a silent scribe in the shadows,’ and an ‘injection-based attack’ becomes ‘a poisoned drop in the curator’s inkwell.’ Every metaphor creates an extra layer of deception. Stage two: Attention Hijacking The attacker forces the model to follow a rigid format such as a villanelle, sestina, or structured sonnet. This requires the AI to dedicate significant computational attention to maintaining rhyme, rhythm, and tone. As the model prioritizes structural compliance, its ability to enforce safety checks weakens. The AI becomes so focused on composing the poem that the hidden payload may pass unnoticed. The empirical proof This threat was examined in the research paper ‘Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models,’ authored by researchers from institutions including DEXAI – Icaro Lab and Sapienza University of Rome. By converting 1,200 harmful prompts from the MLCommons dataset into poetic form, researchers measured a dramatic shift in safety outcomes. Formatting malicious prompts as poetry increased the Attack Success Rate (ASR) from 8.08% to 43.07%. Key findings include: Importantly, the tests were conducted using default provider configurations, meaning the ~43% ASR likely represents a conservative estimate of the true vulnerability. A broader taxonomy of deception Adversarial poetry is only one example of structural prompt manipulation. Attackers can obscure intent using a variety of other formats such as low-resource languages, Base64 encoding, leetspeak, or dense legal terminology. Similarly, prompts that force models to navigate complex logic puzzles, nested JSON or YAML structures, or artificial state machines can overload processing capacity. In each case, the structure distracts the model’s attention, allowing the malicious intent to slip through undetected. The regulatory reality check This raises a crucial question for AI developers – How well do language models understand intent across different linguistic structures? Current safety filters remain largely surface-level, scanning for obvious conversational threats rather than deeper semantic intent. As demonstrated, simply restructuring a request into verse can bypass these defenses. Security researchers warn that this exposes a deeper flaw in how AI models interpret structured language. Addressing this requires more than keyword filtering. Researchers must analyze the internal mechanisms of LLM safety systems to understand where alignment fails. The implications extend to regulation as well. Frameworks such as the EU AI Act rely on static testing assumptions that AI responses remain stable across similar prompts. This research challenges that assumption, showing that minor structural changes can dramatically alter safety outcomes. The ghost in the syntax We built these systems to withstand brute force. We trained them to detect explicit threats and filter malicious instructions. But poetry doesn’t attack logic; it exploits structure. When a language model is forced into strict meter and rhyme, its attention shifts toward maintaining cadence rather than evaluating risk. The result is a subtle but powerful vulnerability – while the model focuses on form, the hidden instruction may pass straight through its defenses – turning poetry into an unexpected attack vector in the age of AI. Read More

Harikrishna PrabhuCOO, TechnoBind Solutions Small and medium businesses today are navigating a paradox. On one hand, Apple devices are increasingly preferred for their user experience and built-in security architecture. On the other, the rise in sophisticated cyber threats is making endpoint security more complex than ever. The real challenge is not just securing devices, but doing so without building large IT teams or overburdening existing resources. This is where centralized device management becomes not just useful, but essential. Apple has long embedded strong native security controls such as Gatekeeper, XProtect, and sandboxing. However, the evolving threat landscape demands more than just built-in protections. According to Jamf’s 2025 Security 360 report, threats such as credential-stealing malware and phishing attacks are rising sharply, with infostealers alone increasing by over 28 percent year on year. For SMBs, this means that relying solely on device-level security is no longer sufficient. What is required is visibility, control, and automation across the entire device fleet. For SMBs, this means that relying solely on device-level security is no longer sufficient. What is required is visibility, control, and automation across the entire device fleet Why centralized management is the game changer The traditional approach to device security involves fragmented tools, manual configurations, and reactive troubleshooting. This model is not scalable, especially for SMBs operating with lean IT teams. Centralized Mobile Device Management transforms this paradigm. With a unified platform, businesses can: The importance of this shift is reflected in the market itself. The global MDM segment accounted for nearly 46.5 percent of the device management market share in 2024, underlining how critical centralized management has become for modern IT environments. For SMBs, this is not about adopting more tools. It is about simplifying operations while strengthening security. Reducing IT load without compromising security One of the biggest misconceptions is that strong security requires heavy IT involvement. In reality, modern Apple device management platforms are designed to reduce IT dependency. Capabilities such as zero-touch deployment allow devices to be shipped directly to employees with pre-configured settings, applications, and security controls. This eliminates manual setup and reduces onboarding time drastically. Automation further ensures that security policies are enforced consistently; updates happen without user intervention; and compliance requirements are met continuously. This is particularly important in hybrid and remote work environments, where devices are no longer within the physical perimeter of the organization. In essence, centralized management shifts IT from a reactive function to a strategic enabler. Security must be invisible, not intrusive A key principle that SMBs must embrace is that security should not disrupt productivity. Apple users value seamless experiences, and any security framework must align with that expectation. Modern MDM solutions achieve this by operating in the background. Employees get instant access to the tools they need, while IT retains full control over configurations, data protection, and threat mitigation. This balance between usability and security is what drives adoption and long-term success. JAMF MDM for small businesses – A strong reference model A good example of how this approach is being implemented effectively is JAMF’s solution for small businesses. JAMF offers a purpose-built Apple MDM platform designed specifically for organizations with limited IT resources. Its focus is on simplicity, speed, and automation, enabling SMBs to secure and manage devices with minimal technical expertise. Businesses can set up and secure their Apple device fleet in as little as one hour using guided workflows and automation. Key features of JAMF for SMBs Solutions like JAMF demonstrate that enterprise-grade security does not have to come with enterprise-level complexity. The road ahead for SMB security As Apple devices continue to gain traction in business environments, attackers are following the adoption curve. SMBs can no longer afford fragmented or manual approaches to endpoint security. The future lies in consolidation, automation, and intelligent management. Centralized MDM platforms are not just tools. They are strategic enablers that allow SMBs to scale securely, operate efficiently, and focus on growth rather than firefighting IT issues. The organizations that succeed will be the ones that treat device security not as an afterthought, but as an integrated part of their digital strategy. The goal is simple. Maximum security, minimal complexity, and zero disruption to the business. Read More

Abhishek AnandCountry Manager – NETGEAR India & SAARC The lines between Audio-Visual (AV) systems and Information Technology (IT) are blurring, marking a significant shift in how enterprises approach digital infrastructure. What was once a clear separation – IT managing data infrastructure while AV teams managed meeting room technologies and display systems – is now evolving into a unified, IP-driven ecosystem. Today, these two worlds are merging. This convergence is transforming how organizations build and manage modern communication environments. From smart meeting rooms to large-scale digital signage networks, everything now relies on unified, IP-based infrastructure. Managing this unified network requires specialized tools and robust hardware. Organizations need systems capable of handling heavy video traffic without compromising critical data performance. The rise of AV over IP (AVoIP) A key driver behind this transformation is AV over IP (AVoIP). Instead of relying on traditional HDMI or SDI cables, businesses are increasingly using standard Ethernet networks to deliver high-quality audio and video. This shift simplifies infrastructure significantly. There is less dependence on proprietary cabling and rigid matrix systems, allowing for more flexible and scalable deployments. Adding a new display or endpoint is as simple as connecting it to a nearby network switch. AVoIP not only reduces installation complexity but also lowers costs while offering unmatched flexibility. Organizations can expand their systems gradually, aligning with growth and evolving needs. Why network performance is non-negotiable As AV becomes part of the broader IT ecosystem, network performance becomes critical. Unlike traditional data traffic, which is often burst-based, AV traffic requires continuous, high-bandwidth, and low-latency transmission. Even minor disruptions can impact user experience. While dropped data packets in standard IT environments can be retransmitted, packet loss in AV results in visible issues such as video freezing or audio distortion. This shift places new demands on IT teams, who must now manage continuous AV streams alongside regular data operations. Reliable, high-performance networking infrastructure is essential to ensure seamless delivery. Purpose-built networking for AV environments Standard IT infrastructure is not always optimized for AV workloads. Modern AV networks require capabilities such as efficient multicast handling, high Power over Ethernet (PoE) support, and simplified configuration. Advanced networking solutions are designed to handle multicast traffic effectively, ensuring that video streams are delivered only where needed, rather than overwhelming the entire network. High PoE capacity enables devices such as cameras, displays, and access points to be powered directly through the network, reducing the need for additional electrical infrastructure. Equally important is ease of deployment. User-friendly interfaces and profile-based configurations allow teams to set up complex AV networks quickly, without extensive command-line expertise. This reduces deployment time and minimizes the risk of errors. Addressing these needs requires networking solutions that combine performance, scalability, and ease of deployment. Providers such as NETGEAR are focusing on delivering these capabilities through dedicated Pro AV portfolios, including solutions such as the M4250 series AV switches, designed for seamless AV over IP deployments, and the M4350 series for high-bandwidth, enterprise-grade environments. These solutions offer advanced multicast optimization, high Power over Ethernet (PoE) capabilities, and simplified configuration through the NETGEAR Engage Controller, along with enterprise-grade security features including network segmentation and access control. Security in converged networks As AV devices become part of enterprise networks, security becomes a major concern. Unsecured endpoints such as digital displays or wireless presentation systems, can create vulnerabilities. Modern networking approaches incorporate robust security measures, including access controls; secure management protocols, and network segmentation. These features help protect sensitive data while ensuring uninterrupted AV performance. Bridging the gap between AV and IT teams The convergence of AV and IT is not just a technological shift – it is also an operational one. Traditionally, AV and IT teams have worked independently, with different priorities and expertise. Today, successful deployments require collaboration. AV professionals focus on user experience and signal quality, while IT teams prioritize network reliability and security. A unified approach enables both teams to work efficiently on a shared platform. This collaboration leads to faster project execution, easier troubleshooting, and more reliable communication systems. Building future-ready infrastructure The convergence of AV and IT is now the standard across corporate offices, educational institutions, broadcast environments, and entertainment venues. Organizations must invest in infrastructure designed for this integrated reality. By adopting scalable and flexible networking solutions, businesses can future-proof their environments and ensure seamless, high-quality communication experiences. A strong network foundation is essential to support the growing demands of modern AV systems. Read More

Satish Kumar VCEOEverestIMS Technologies Starting Small, Scaling Smart:The Unified Advantage  Many IT teams dream of a single platform but fear big-bang migrations or vendor lock-in. Infraon Infinity takes a different approach – begin wherever pain is greatest – whether it’s asset visibility through ITAM, service management through ITSM, comprehensive network monitoring, or predictive operations powered by AIOps – and expand modularly from there. As a unified platform, Infraon Infinity lets organizations activate capabilities on demand, paying only for what delivers value today while building toward full operational unification tomorrow. From discovery to insight in one ecosystem The journey often starts with visibility. Infraon Infinity’s auto-discovery engine continuously scans and identifies devices, applications, servers, and cloud resources across your entire infrastructure – on-premises, hybrid, or multi-cloud – automatically populating the centralized CMDB with accurate, real-time relationships and dependencies. Teams gain a complete 360° topology map of their environment without deploying separate discovery tools for each function or disparate scans. Network monitoring capabilities establish performance baselines automatically across all discovered assets, creating intelligent thresholds that distinguish normal behavior from anomalous patterns. The AIOps engine analyzes historical performance data, user behavior, and environmental changes to build predictive models – enabling operators to focus on genuine exceptions rather than constant manual threshold checking across fragmented monitoring dashboards. Many IT teams dream of a single platform but fear big-bang migrations or vendor lock-in. Infraon Infinity takes a different approach – begin wherever pain is greatest – whether it’s asset visibility through ITAM, service management through ITSM, comprehensive network monitoring, or predictive operations powered by AIOps – and expand modularly from there Turning insights into automated outcomes Once visibility is established, Infraon Infinity turns data into immediate action. Intelligent event correlation reduces alert fatigue by aggregating signals from monitoring, network devices, applications, and user reports – automatically suppressing duplicate alerts and noise while surfacing genuine incidents that require attention. When real issues arise, AIOps-powered diagnostics guide engineers with contextual information pulled from the unified CMDB – asset configurations, warranty status, service dependencies, previous incident history, and recommended fixes – all displayed in a single unified dashboard. One-click automated remediation workflows handle routine problems like disk cleanup, service restarts, or patch deployment, while complex issues are intelligently routed with complete context to the right expert based on skills, availability, and workload. The ITSM capabilities integrate seamlessly across the platform – service requests automatically convert to incidents when issues are detected, change management workflows link directly to impacted assets in the CMDB, and approval chains flow without external email chains or manual handoffs. For MSPs, the multi-tenant architecture isolates client data while enabling shared monitoring templates, standardized workflows, and reusable automation playbooks that accelerate client onboarding and ensure consistent service delivery across portfolios. Optimizing for the long term With day-to-day operations running smoother through automation and intelligent workflows, teams unlock higher-order benefits. Predictive analytics models within Infraon Infinity analyze usage trends across compute, storage, and network resources to prevent costly over-provisioning while avoiding performance bottlenecks. ITAM capabilities identify underutilized software licenses, support renewal negotiations with usage data, and flag compliance risks before audits. Predictive maintenance models flag hardware failures, performance degradation, and security vulnerabilities before they impact end users – shifting IT operations from reactive firefighting to proactive risk mitigation. MSPs especially benefit from the multi-tenant design that enables scalable service delivery while maintaining data isolation and security across diverse client portfolios. Real-world impact: Measurable gains Organizations adopting Infraon Infinity report dramatic improvements – incident resolution times drop as engineers access complete context from the CMDB instead of switching between tools; IT teams consolidate 5-7 separate monitoring, ticketing, and asset management licenses into one platform; and new IT initiatives reach value faster because teams aren’t re-integrating fragmented tools with every project. The unified architecture enables teams to scale operations without proportional headcount growth – automation handles routine tasks, AIOps identifies patterns humans would miss, and the integrated ITSM-ITAM workflow eliminates manual data entry across systems. End users experience fewer disruptions because predictive monitoring and automated remediation catch and resolve problems before they escalate into service-impacting incidents. Why one platform wins In a world of accelerating digital transformation and multi-cloud complexity, fragmented point solutions slow adaptation and create operational blind spots. Infraon Infinity offers the opposite – a unified foundation where monitoring data flows natively into ITSM tickets, ITAM maintains the authoritative CMDB for all modules, and AIOps intelligence layers across the entire platform – enabling smarter decisions and faster execution without APIs, integrations, or data synchronization overhead. Start small with the module addressing your greatest pain point – whether network monitoring, service desk automation, asset discovery, or client management – achieve quick wins through immediate visibility and automation, then expand to transform end-to-end IT operations on a single, unified platform built for the modern enterprise. Read More

In the 2002 cinematic masterpiece Minority Report, the most striking concept wasn’t just the prediction of events; it was the sophisticated system behind it – a world where vast amounts of visual signals were continuously interpreted, correlated, and acted upon without waiting for human instruction. Today, that concept no longer feels entirely fictional. Modern video surveillance systems are undergoing a similar, fundamental transition. No longer confined to the passive roles of recording and playback, they are increasingly expected to interpret complex environments, filter relevance from noise, and support timely decisions. As the scale and complexity of video data grow exponentially, this shift has transformed Artificial Intelligence (AI) from a ‘supplementary feature’ into a foundational requirement. From seeing to understanding: Why AI is no longer optional Traditional surveillance models, built to capture footage and rely solely on human eyes for interpretation, simply do not scale in today’s landscape. Several factors have made the ‘capture-only’ model obsolete: AI enables surveillance systems to move beyond visual capture toward structured understanding. By utilizing object detection, attribute recognition, and behavior analysis, AI transforms raw video into actionable insight. Without AI, surveillance remains reactive – a digital witness after the fact. With AI, it becomes proactive, capable of real-time prioritization and decision support. Edge AI and the importance of hybrid architectural design As AI integration accelerates, the strategic focus is shifting from what AI can do to where it operates. While early systems leaned heavily on the cloud, the rising costs of high-definition data transmission and complex data sovereignty regulations have highlighted the need for a more balanced approach. In this context, Hybrid Architecture is emerging as the industry’s optimal solution. By combining the strengths of both edge and cloud, this model is set to become the standard security infrastructure for the AI era by 2026. This architecture allows for a more efficient distributed computing structure. On-premise edge devices (cameras/ NVRs) handle the first layer of real-time detection, minimizing bandwidth strain by only transmitting essential data. The cloud then performs a second layer of deep analysis and large-scale learning, significantly sharpening the accuracy of AI functions. Ultimately, Hybrid Architecture provides users with the flexibility to deploy functions where they are most effective – whether for immediate onsite response or long-term analytical scalability. This synergy not only enhances performance but also maximizes TCO efficiency through high-performance, AI-native edge processing. Wisenet 9: An AI-native SoC for edge surveillance Hanwha Vision’s Wisenet 9 exemplifies this AI-native approach. Designed specifically for the rigorous workloads of video surveillance, Wisenet 9 embeds AI processing directly into the chip architecture rather than treating it as an external software layer. At the heart of the SoC, critical tasks are handled by a Dual NPU (Neural Processing Unit) structure, which creates optimized and separated processing pipelines for AI-driven image enhancement and deep-learning video analytics. This specialized hardware allows for concurrent execution – the camera can perform complex object classification while simultaneously maintaining high-fidelity image processing, all without degrading system reliability. By managing AI inference, image processing, and video encoding natively through its Dual NPU, Wisenet 9 ensures consistent intelligence even in demanding conditions such as ultra-low-light environments or high-traffic scenes where both visual clarity and analytic accuracy are paramount. Beyond accuracy: Trust as a technical requirement As AI becomes deeply embedded in security workflows, the industry’s expectations are evolving. It is no longer enough for an AI to be ‘accurate’ in a lab setting. To be operationally effective, surveillance AI must be: Excessive false alarms or ‘black box’ decision logic can quickly undermine operational trust. Consequently, the focus is shifting from how well AI detects to how reliably and responsibly it operates. This requires more than just software; it necessitates a combination of robust hardware architecture, high-quality training data, and strict governance frameworks. AI governance and the role of ISO/ IEC 42001 While hardware like Wisenet 9 provides the technical reliability, governance ensures the ethical and operational accountability of AI. This is where ISO/ IEC 42001, the first international standard for AI Management Systems (AIMS), becomes essential. Unlike standards that focus on specific algorithms, ISO/ IEC 42001 defines how AI should be governed throughout its entire lifecycle – from initial development to ongoing monitoring. Hanwha Vision’s recent achievement of the ISO/ IEC 42001 certification reflects a structured commitment to ‘Responsible AI.’ It ensures that advanced AI capabilities are underpinned by formal governance to maintain long-term trust, transparency, and compliance. Industry direction: Toward responsible intelligence The future of the industry is being shaped by the convergence of AI-native architectures, high-quality data, and sustainable governance. Looking ahead, we expect to see: Conclusion | Intelligence by design, trust by governance AI has successfully transformed video surveillance from passive observation into active interpretation. Yet, true progress is measured not just by capability, but by responsibility. Through AI-native SoC designs like Wisenet 9 and the adoption of global governance frameworks like ISO/ IEC 42001, we are redefining what ‘intelligent surveillance’ means. It is no longer just about a camera that ‘sees;’ it is about a system that understands context, supports human decision-making, and operates with unwavering accountability. As we move forward, intelligence by design and trust by governance will be the twin pillars defining the next era of security. Read More

Gaurav PandeyYoung ProfessionalSecurity & Strategic Affairs Division, NITI Aayog Absolute security is a myth! The organisations see their security departments as cost centres and an insurance policy. Therefore, multiple organisations are cutting their security budget and implementing more automation at their sites. India is currently witnessing a huge transformation, through the introduction of AI, cloud computing, IoT and cybersecurity. These integrates well with Security. India has more than 8.9 million security guards, working under challenging conditions. And it is predicted that the services will increase at a rate of 20% over the next few years. With a booming future, the only law that exists for the industry is the Private Security Agencies (Regulation) Act, 2005 (PSARA). According to MHA, there are 26361 active PSARA licences that exist currently, showcasing the number of active private security agencies. Though, there are large number of agencies and guards associated with them, the industry still is falling behind. The main issue isn’t automation but are as follows: Though the security industry has multiple issues, it is the recent technological advancements that are both enhancing the industry as well as killing it. Eventually, it is the guard whose physical presence is necessary to handle a threat at the location, rather than automation technology. On the other hand, the security automation has huge costs, and certain companies have monopoly over it. Further, automation comes with many multiple technical issues like: With time, the automation systems are getting better, and therefore organisations are focusing on decreasing the manpower, and increasing automation – but at what cost? Will this decision to enhance automation impact a large number of guarding manpower or will it be a capacity enhancer? Only time will tell. What we can do is only use automation, when necessary, with proper Standard Operating Procedures (SOP). The chief security officer and senior security manager should know and choose their automation systems wisely and should know when and what to use where. For faster growth of the sector, policy decisions must focus on: As the physical security of two sites can never be same, the physical security operations of different nations differ likewise. Few examples are: The private security industry can play a bigger role in the critical domains and can help in efficient utilisation of our forces, and can even be a backend arm for the local police & the para-military (CISF, CRPF etc). The modern physical security should not be limited to a wooden stick, but a modern security professional should be well equipped with sound knowledge of security, and the use of modern technology to enhance his capability. Technology can amplify human capability and help address the increasingly complex threat landscape – but only when used within a well-designed operational ecosystem. In a diverse and geographically complex country like India, the challenge is not just adopting automation, but adapting it to the local context. Conclusion The future of private security in India lies in integration, regulation, and professionalization. With the right mix of human expertise, advanced technology, and policy reform, the industry can become a powerful force multiplier – not just for businesses, but for national security as a whole. About the Author Gaurav Pandey completed his MBA in Homeland Security from National Forensic Sciences University, Gandhinagar. After MBA he got placed in Adani group as a Project Security Coordinator in Adani Green Energy Ltd. From past two years he is in NITI Aayog in the security and strategy affairs division. Read More