An AI research agenda for and by the telecommunications industry

Artificial Intelligence (AI) is a transformational and transversal technology, applicable to any economic sector and many aspects of our lives. With the appearance of ChatGPT at the end of 2022, the awareness of AI has grown exponentially. Consequently, in 2023, many reports have appeared about the impact, trends, applications, economic impact, and risks of AI, both across industries and industry-specific. Most of those reports are written by analyst or investor companies such as IDC, PWC, Mckinsey, Banc of America, etc.

The main areas of AI research by and for the telecommunications industry.
Richard Benjamins

Richard Benjamins Follow

Reading time: 4 min

In the fall of 2023, a different type of report has appeared about the use of AI in the telecommunications industry, namely about the research needs in AI tailored to the telecommunications industry. The main difference with other reports is that this one is conceived and written for and by the industry.

In November 2022, the GSMA, ETNO, Telefonica and the Humane AI Net project (funded by the European Commission) organized a workshop in Munich, Germany dedicated to the research needs in AI of the industry. The operators that participated included Axiata, O2, Orange, STC, Telefonica, Telenor, Telia, Telstra, TIM, Turkcell and Vodafone. After an introduction of the latest AI research trends by the Humane AI Net partners and an overview of upcoming AI regulations by ETNO, the operators started discussing on current and future uses of AI. This discussion has been the basis for the “AI Research agenda for the telecommunications industry” which you can access here.

The figure shows the table of content of the report.

AI Research agenda for the telecommunications industry

Data Foundations

Data foundations are important for AI because they provide the raw material that AI algorithms use to learn and make predictions. Without access to high-quality, diverse, and relevant data, AI systems would not be able to perform their tasks effectively. Data foundations also play a critical role to ensure the scalability and maintainability of AI systems over time. Relevant topics include privacy, data anonymisation and synthetic data.

Scaling AI

Most operators have started to use AI to improve their businesses in several different respects, as we have seen. However, a remaining challenge is how to scale the use of AI to every corner of the business: optimisation, operation, marketing, customer interaction, new products and services, new business opportunities, and horizontally in digital transformation processes. Topics of interest include standardisation of “commodity” telecommunications use cases and AI industrialisation with MLOps.

AI applied to the network

The telecommunications industry is already using AI to improve its core infrastructure, the network, in several ways. But there are still many areas of improvements that require more research before they can be applied at scale, including 5G core and RAN, near real-time optimisation of network, network automation, anomaly detection, explainable AI, digital twins and Naas (network as a service).

Operations and marketing

Optimizations of operations is one of the most popular areas to apply AI, because it allows for large savings and increase efficiency. Typical areas where AI is already applied are: next-best activity (NBA), churn prediction, smart pricing, credit scoring,       device recommendation, product and service recommendation, and digital assistants chatbots. Moreover, this area of application is similar for many sectors. Topics that still require research include real-time AI and the combination of optimization and machine learning to make it possible to handle more complex and dynamic optimisation problems.

Customer interaction (chatbots and virtual assistants)

Chatbots and virtual assistants are helping companies to interact 24/7, 365 with their customers in a real-time and personalised manner. Current research topics that need further investigation include proactivity in the interaction and better dialogue capacities. More proactivity allows anticipating the needs of the user and provide relevant information or perform tasks without the user having to specifically ask for it. It is, however, important to find the right balance between personalisation and generating the impression of “spamming” the user.  Better dialogues imply to move from question-answering to having conversations, something the LLMs are enabling. Moreover, multi-language chatbots are important for multinational operators.

Responsible business and ethical AI

With the massive adoption of AI, its ethical use and societal impact is increasingly important. Topics that still require research include the ability to perform a homogeneous risk qualification; to perform this across the value chain with providers, open source and partners; ethical AI tools for bias and explainability; green AI and AI for Good.

B2B/B2G data sharing and the data economy

Companies are holding vast amounts of data, which are mostly used to benefit the businesses of those companies. There is, however, also a huge opportunity for data sharing between companies, within and across sectors. This is also referred to as the data economy, an upcoming economy that is still in an incipient state. Research topic in this area include standardisation and interoperability of data sets, trust and sovereignty, privacy of personal data, federated data sharing and federated ML, and assurance of ethical use.

Aditional research topics

Finally, some additional research topics relevant for the telco industry include AI as a Service (AIaaS), the metaverse, and quantum computing.

The research agenda can be accessed through this QR code.


Contact our communication department or requests additional material.

Telefónica Centenary logo Celebrate with us the Telefónica Centenary