Telco Talks with Terje Jenson - Beyond Detection Speed: The Real Business Value of AI-Driven Security

Beyond detection speed, what are the most significant operational or business benefits you've observed since implementing AI-driven security capabilities?

Telenor business security spans a 360-degree perspective including threat intel, reactive, proactive and collaboration/scaling topics. AI is used in several of these areas besides the ability to swiftly detect, understand and respond to an incident. Benefits include the ability to assess huge amount of data from different sources for example when assessing threats and translating these into proper actions tailored to each business scope. Another use of AI is to interact with the document base to lower thresholds for anyone to follow the best security designs. AI can be used to do an automated assessment of actual configurations. It can also be extended to gamified training and awareness for competency building as well as strengthening the overall resilience of the organization.

What is the biggest challenge you have faced in scaling AI capabilities across your security operations and how are you addressing it?

For several legacy systems, timely access to data with proper quality is a challenge. One additional complication comes when data from several systems with different formats and interpretations have to be consolidated. One way to addressed this is to clean up the data, translate and re-format adopters, use master data base, or similar.

Another area to consider is privacy and confidentiality of information, which implies that different data sources are properly labelled and categorised. This may be relevant when external AI models are used where information might be shared outside the trusted enterprise scope.

There are different AI implementations such as standalone and embedded in other systems. This could increase the complication when there is a need to fully control the training and inference phases. Inventory of the different main AI usages and AI agents could be required, which place a task on consistently describing their profiles.

When it comes to trust, governance, and accountability in AI-driven security systems, what frameworks or principles guide your decision-making?

There are several frameworks and sets of principles applicable for AI – such as responsible AI and Cloud and AI directives, AI agent frameworks and similar. These are inputs when shaping Telenor’s practical AI governance. There are several use cases and corresponding risk scopes for AI. In many cases AI is used for improving personal productivity. While is important that governance principles are followed, it is also essential to keep the level of bureaucracy on a pragmatic level. For example, the risks could be different for preparing a public presentation compared to is using AI in a customer chatbot.

Several bodies recommend having inventories of the most important AI models and implementations. These could span use case such as customer interactions, network optimization, product and system development and employee processes. Having a risk-based approach with perspective on value-at risk would be central to guide on proper guardrails to apply.

What are you looking forward to most at Network X this year? And what do you hope attendees take away from your session?

Network X is a good opportunity for meeting new and old colleagues and engaging in talks on common challenges and ways to solve these. The mixed formats with presentations, panels, demos and social mingling facilitate such interactions. Persons from several regions and company types are present, which represent broader perspectives on new and innovative ways to grow and improve.

Topics like 5G – preparing for 6G, cloud and broad band represent large investment areas across the globe. Hence, it is essential to have a solid understanding on how to act on these in effective and robust ways that also give business opportunities. Some examples could be approaches for delivering sovereign and edge cloud, driving use cases for AI factory, maximizing value from autonomous networks.