The Zhitong Finance App learned that on January 24, Omdia published an article stating that generative artificial intelligence is expected to benefit telecom operators in many ways. This includes increased operational efficiency and productivity, as well as associated cost reductions. Generative artificial intelligence can be used to enhance the customer experience (such as personalized customer service and products/packages). It can also improve internal applications (such as document aggregation) and enhance external commercial applications (for a wide range of fields such as e-commerce and financial services, education, healthcare, smart homes, and cities).

Generative artificial intelligence is being integrated into mass-market products and services, and penetrating all major vertical industries, including telecommunications. The impact of generative artificial intelligence is far-reaching and far-reaching. According to Omdia's generative AI software market forecast, global software revenue will grow from $6.2 billion in 2023 to $58.5 billion in 2028, at a CAGR of 56%. The telecom industry isn't the largest vertical in the market (the consumer industry is in first place, followed by media and entertainment), but it is an important industry, and generative software revenue is expected to exceed $3 billion by 2028.

However, generative artificial intelligence is highly disruptive: it can have a negative impact on job functions, and the generated results may infringe copyright. Generative AI faces many ethical and moral issues, including models and outputs that may be biased and discriminatory. Generative artificial intelligence can also make mistakes or be made up indiscriminately. Generative artificial intelligence is a focus of complex, evolving, and unbalanced regulatory governance frameworks. The technology itself is also developing rapidly, bringing a new round of beneficial innovation while also bringing risks and long-term challenges. There are many issues that need to be addressed, but telecom operators must face it because ignoring generative artificial intelligence is not a good idea, and as mentioned above, careful, thoughtful deployment can pay off.

Omdia believes that well-managed data assets are critical to generative artificial intelligence, and most telecom operators aren't doing enough in this regard. Telecom operators have large, fine-grained data sets, especially large carriers. However, due to poor data governance, the advantages often stop there. According to Omdia's 2023 Artificial Intelligence Market Maturity Survey, only 40% of telecom respondents have developed data management and governance plans.

For most telecom operators, building generative AI models from scratch isn't a good idea. Building a model in-house from scratch provides maximum control, but it's a difficult task requiring deep multidisciplinary AI expertise, huge budgets, and resources. Telecom operators in Asia are leading the way in this regard: China Mobile was one of the first telecom operators to try it out. It has successfully created a complete ecosystem (basic model, core infrastructure, generative artificial intelligence platform, and applications).

Fine-tuning the existing underlying model is a more viable option. This process allows telecom operators to adapt pre-built, trained models to suit specific fields and tasks. Fine-tuning requires fewer resources, and there are more and more basic models to choose from.

Telecom operators don't need to build generative AI development tools. There are plenty of open source generative artificial intelligence tools for telecom operators to choose from. Maintaining and expanding proprietary tools and libraries is challenging, and can rapidly deplete telecom operators' artificial intelligence resources.

Telecom operators have infrastructure assets that can support generative artificial intelligence. Good generative AI infrastructure requires high-bandwidth, low-latency network infrastructure, high-speed data transfer protocols, and high-capacity memory standards.

Openness and interoperability are key to telecom operators' success. Generative AI services and solutions are born in multi-cloud and hybrid cloud environments. Telecom operators must avoid siloed architectures and proprietary implementations when launching new applications.

In the age of generative artificial intelligence, the key players in the telecom supply chain are different. Compared with traditional telecom network infrastructure, AI chip vendors, cloud infrastructure, and hyperscale technology companies dominate the generative AI core infrastructure supply chain. The emergence of a new wave of solution providers in the telecommunications sector requires telecom operators to cooperate with more diverse suppliers and establish new partnerships and alliances.