AI-generated visual, Microsoft Designer
Unsurprisingly, the common theme of the Mobile World Congress (MWC), the high mass of mobile which is being held this week in Barcelona, is contained in two letters: AI. While smartphone sales are struggling to recover on the European continent, the “AI inside” stamp could make consumers want to change their terminal. Writing messages, summarizing web pages, automatic photo editing, real-time translation of voice calls… The contributions of generative AI are such that it should constitute, at the time of choice, a discriminating factor between mobiles “Gen AI” and the others.
According to the research firm Counterpoint, the share of smartphones boosted by artificial intelligence should represent 8% of the market in 2024 and reach 40% in 2027, or 500 million units sold. As usual, Samsung fired first. After launching the fashion for “phablets” or more recently foldable smartphones, the South Korean manufacturer presented, with the Galaxy S24 range, the first “smart” mobiles. Its Chinese competitors Xiaomi, Honor and Vivo must follow suit according to the announcements at the show.
Telecom operators also intend to take the AI turn. Deutsche Telekom has unveiled a smartphone prototype whose AI-based assistant replaces many apps. Developed by Qualcomm and the American startup Brain.ai, this T-Phone responds to requests by voice command. The user can use it to go shopping, plan a trip, book a flight or create a video.
Two new alliances around AI
But AI is not just nesting in terminals. Operators are also considering using AI to automate their customer service, optimize the performance of their networks or move to predictive maintenance of their infrastructures. Orange, for example, uses deep learning to find the best location of its relay antennas for optimal mobile coverage.
Opportunistically, certain players could take advantage of the introduction of AI to reduce their payroll. Last May, British Telecom (BT) indicated that the automation of the processing of telephone calls from its customer service or the diagnosis of network anomalies could lead to the elimination of the equivalent of 10,000 positions.
Operators also intend to gain independence from the American leaders in generative AI such as OpenAI/Microsoft, Meta or Google. Within the “Global Telco AI Alliance”, Deutsche Telekom, again, announced the creation of a joint venture with the Emirati operators e& Group and Asian operators SoftBank, Singtel and SK Telecom to develop large language models (LLM ) specifically adapted to telecoms companies and more particularly chatbots and digital assistants dedicated to customer relations.
Also announced at the MWC, the creation of the “AI-RAN alliance” aims to use AI to improve the performance of mobile networks but also to reduce the energy consumption of infrastructures in 5G and tomorrow in 6G. The consortium brings together operators (T-Mobile, SoftBank), equipment manufacturers (Nokia, Ericsson), chip manufacturers (Arm, Nvidia) and public cloud giants (Microsoft, AWS).
An AI to fight against telephone scams
Like last year, these hyperscalers are increasing the number of announcements relating to services specifically dedicated to the world of telecoms. With Copilot in Azure Operator Insights, Microsoft puts its famous intelligent assistant at the service of operators’ technical teams to facilitate the detection and resolution of network problems.
More original, Azure Operator Call Protection aims to fight against fraudulent calls and telephone scams. This new service, which works on any terminal, mobile or fixed, analyzes conversations in real time and alerts the user during a call if it detects suspicious behavior such as, for example, the request for sensitive personal data . British Telecom is using it in preview.
For its part, AWS is increasing its partnerships. The world’s number one public cloud partner has partnered with specialist publisher Amdocs to develop an AI model responding to major use cases from network orchestration and operation to customer engagement. Based on Claude, Anthropic’s generative AI, the solution, called Amdocs Intelligent OSS, is hosted on the Amazon Bedrock platform.
AWS also collaborated with Snowflake, publisher of a data cloud platform, to create GenTwin. This solution dedicated to network operations combines private 5G, generative AI via Amazon Bedrock and the digital twin concept using AWS IoT TwinMaker. Network virtualization allows for greater automation and scalability.
“Telecoms AI” is upon us
Finally, note two initiatives, from Qualcomm and Red Hat, in terms of AI and telecommunications.
Qualcomm indeed presented its 13 GHz GIGA-MIMO technology at MWC24. This technology helps improve the efficiency of AI services as 5G Advanced is deployed. MIMO is an antenna technology that increases wireless communication capacity. Qualcomm uses these multiple antennas on base stations and terminals to increase capacity through the number of antennas. This means that the average transmission capacity increases with the number of base stations and terminals.
Finally, Red Hat and Japanese telecommunications company NTT announce a collaboration with Nvidia and Fujitsu to develop real-time generative AI data analysis solutions on edge computing, as part of the Council’s IOWN initiative next generation communications standards. The solution is built on Red Hat OpenShift on top of Kubernetes using technology developed by IOWN.