AI调查报告

SURVEY REPORT AI Survey 2025 Introduction As an industry based on innovation and progress, mobileis always looking for the ‘next big thing’. In the past,there have been many candidates. From VR to foldablesto metaverse, they have come and gone with varyingdegrees of success. But can there be any ‘next big thing’ to rival ArtificialIntelligence (AI)? It’s fair to say AI is now sweeping across the world,promising to transform virtually every industry vertical.The trigger for this is undoubtedly AI’s new ‘flavour’,generative AI. Large language tools such as ChatGPTand Claude have revealed AI’s ability not just to detectpatterns and learn skills, but to generate original content.In so doing, they have moved GenAI tech into themainstream. Indeed, the clamour has reached a point where,according to PwC’s Global Artificial Intelligence Study,AI is predicted to add$15.7 trillionto the globaleconomy by 2030. Needless to say, mobile is also in thrall to thistechnological revolution. Many telco insiders will counterthat this ‘revolution’ is nothing new. They will argue thatthe mobile industry first deployed AI (or more correctly,machine learning) decades ago to automate everythingfrom network planning and maintenance to virtualagent-based customer care. They are right, of course. In fact, as far back as 2017,the GSMA brought together mobile operators andpartners to create an AI for Impact initiative. It workedwith private and public sector organisations to use bigdata for social good and to ensure the mobile industryis deploying AI responsibly. But it’s still true that something has changed in the lasttwo years. This tech is now dominating the conversation,pushing AI chip designer Nvidia to a market cap of$3.3trillionand giving AI’s most famous startup, Open AI,a$157 billionvaluation. In the mobile space, analysts have been busy calculatingthe possible impact of the new wave of AI innovation.McKinsey, for example, believes revenue gains from theexpanded use of the tech within the telecoms sectorcould be$680 billionover the next 15 to 20 years. This shouldn’t surprise us. One way to think aboutAI is that it does for brain power what the industrialrevolution did for horse power. This opens the wayfor unprecedented productivity gains. It’s important to note that the new AI processes aredifferent from traditional rule-based and determinedalgorithms. True AI should include a significant elementof learning from training data in order to make decisionsand perform cognitive functions. So how do mobile professionals feel this transformationwill impact their organisations? What are their priorities?Do they have the budget and the skills to deliver RoI? At Mobile World Live, we decided to find out. We werepersuaded by the results of our first ever industry surveyof C-level execs, which we published in 2024. In it,we asked which new business area was the most attractive.The answer ‘advanced AI’ came out top with a 30 per centshare. That was 2x more than the second option. That settled it. So we launched an ambitious project tocollect the opinions of our readers. Just under 300 respondents completed the survey, ofwhich the three biggest areas of business were SoftwareDevelopment (20%), Mobile Network Operators (18%)and Consultancies. 44% of respondents were eitherC-Level, Directors or VP-Level, with 54% of allrespondents based in Europe, 16% in North Americaand 13% in Asia. This is the result. We hope you find it interesting. Contents ·Managing the network·Keeping the network safe and performing·The commercial advantage of AI·Transforming customer care·Transforming voice services·Improving the user experience Managing the network IN PARTNERSHIP WITH Predict network failures AI tools can scrutinise historical regression andclassification models to detect patterns of failure.They can learn from these patterns to pinpointspecific areas of vulnerability, and alert networkoperators to take proactive measures. For all their interest in value added services andincremental revenue, telcos know that in the end itall comes back to the network; the network underpinseverything. With every new cellular generation, there’sadded complexity. The arrival of software-definedcloud-hosted standalone 5G upped the game again.It raised the amount of data flowing across the networkto unprecedented scale. Allocate resources Once the system has identified an issue, it canautomatically adjust network resources basedon real-time demand. It can adjust parametersto reduce energy usage too. This ‘inflation’ makes the task of managing the networkmore onerous than ever. How can engineers rise to the“beyond human scale” challenge of monitoring andfixing network business functions in real-time, 24/7? Give engineers better information When faced with an incident, engineers can beoverwhelmed with information. How can they sortthrough the documentation and reports to isolate theproblem and act on it? This is where generative AItools come in