Fraud management, cybercrime, ransomware
Security leaders need deep observability to balance innovation and risk
Chime Mazal•
July 16, 2025
Artificial intelligence is embedded in modern business fabrics. From automating decisions to optimizing operations, AI innovation is moving at a transformative pace, revealing key gaps in vision, governance, and control.
reference: Beyond Replication and Versions: Protect S3 Data in the Face of Advanced Ransomware Attacks
Almost 80% of businesses around the world use AI in at least one business function, creating new challenges for their security teams. Security experts are expected to move faster, cover more positions and reduce risk across a rapidly evolving modern hybrid cloud environment.
This World AI Appreciation Day is now time to challenge the assumption that rapid innovation always comes at the expense of security. Security leaders have the opportunity to redefine their security strategies. The goal is to balance AI conversion power with a robust governance model.
To achieve this balance, deep observability must be achieved in all data moving across the organization. Traditional monitoring tools have not been built due to the complexity of modern AI workloads. It was also not designed for the advent of Shadow AI applications.
Organizations can scale their investment in monitoring tools by optimizing with network-derived telemetry. This telemetry is delivered in the form of packets, flows and metadata to help organizations respond to today's AI innovations.
Deep observability combines traditional log data from existing tools with network-derived telemetry. This fusion leads the complete image to focus and provides detailed insight into all data moving across the network. Insights include encrypting traffic and flowing through hybrid and multicloud environments as well as lateral east-west traffic that attackers often hide. Security teams provide full visibility into what's happening across your network and detect threats and anomalies that traditional security tools may be missing.
AI innovation affects
Companies compete to take advantage of AI's benefits, but they are encountering a new set of risks. One risk – the biggest concern – is Shadow AI, including the use of unauthorized tools, employee-sourced models, developer-driven deployments.
AI workloads are driving unprecedented spikes in network traffic volumes. A 2025 Hybrid Cloud Security Survey found that one in three organizations has doubled their traffic volume over the past two years due to the adoption of generative AI.
This surge in traffic creates more complexity and introduces blind spots. Security teams are challenging themselves to respond to these rapid changes.
To realize the amount of time companies want to achieve with AI adoption, they need to build robust organizational policies and governance. This is especially important for large-scale language models. Without these guardrails, AI poses unacceptable business risks, threatens customer trust, and threatens regulatory compliance and long-term security.
New threats in the AI era
Cyber enemies use AI to hone their skills and enhance their attacks. An increase in the number of violations indicates that criminals have been successful in using these tactics.
Violations have increased by 17% over the past year, affecting 55% of organizations around the world. These attacks continue to grow in scale and sophistication as attackers mobilize AI to automate their activities. A majority of security stakeholders -58% – have seen an increase in ransomware attacks powered by AI. Almost half -47% – has increased attacks targeting AI or LLM deployments. The volume of phishing and social engineering attacks, as well as the incidence of AI-driven malware and network exploits, is also rising.
Turning the tide requires full visibility to prevent attackers from hiding their presence in the massive flood of traffic generated by AI workloads.
Security at business speed
The democratization of AI has led to an explosion of experiments across modern hybrid cloud infrastructures. Traditional security strategies and tools were not designed to support rapid business transformation. They certainly weren't built to evolve at the speed of AI.
Many of today's security teams are still thinking from a long-standing or quarterly perspective rather than just a few hours. The key to ensuring innovation is to incorporate CISO into AI framework development from the start.
After the fact, security cannot be bolted to a new operational workflow. You need to integrate it into your design from the beginning.
Balancing security and innovation means closing the visibility gap and streamlining the governance model. Security must also be integrated into the organizational culture. This requires a stronger relationship between the CISO and the board.
Future paths: real-time insights and deep observation
Accuracy is required for safe AI adoption. Today's security strategies require the ability to respond to rapidly changing deployments, increased complexity and increased traffic. Security leaders need to change their mindset. Instead of asking, “How do you catch up?”, they should ask, “How do you lead?”
Transformation starts with embedding security at every stage of AI deployment, from experimentation to production. It also means building a stronger bridge between security, DevOps, and the boardroom. Deep observability supports this alignment by providing practical network intelligence. All stakeholders can understand and trust this intelligence.
This approach explains why 88% of security leaders agree that deep observability is important to ensure AI deployment. On this World AI Appreciation Day, we need to challenge the notion that security must curb innovation.
CISOS allows AI to be secured at business speed with the latest technology that supports your cybersecurity strategy.
Ready to learn more?
Download Gigamon 2025 Hybrid Cloud Security Survey to discover more about how security leaders balance the risks and benefits of AI adoption.
