Posted by Mumbai: India’s daily mulitimodal AI use expected to nearly double to 43% by 2030 reveals new Ericsson ConsumerLab research. India is among the markets expected to see strong growth in multimodal artificial intelligence (AI) usage, as per the research.
Multimodal AI refers to artificial intelligence systems or models that can process and integrate information from multiple modalities or sources of data. These modalities can include text, images, videos, audio, and other forms of sensory data. Google Gemini, OpenAI GPT-4o and Anthropic Claude 3 / 3.5 are examples of multimodel AI models.
India’s daily multimodal AI use
India’s daily multimodal AI use across text, voice, images and video projected to reach 43% by 2030, compared with 22% today and 30% globally. Interestingly, there’s a shift in AI use pattern from odd interactions to more frequent across different devices and locations, creating new demands on mobile networks, as per research findings. The shift in AI use is also is likely to show impact on the network infrastructure.
As consumers increasingly use AI while on the move and share more images, video and voice with AI services, network traffic is expected to become more upload-heavy and unevenly distributed. This will increase the importance of network responsiveness and uplink capacity, particularly in locations where network demand is already high.
“AI is becoming an integral part of how consumers search, create and communicate today. As AI becomes more multimodal and mobile, consumers will expect these experiences to work reliably across devices and wherever they are,” said Nitin Bansal, MD – Ericsson India.
“India’s strong digital foundation provides an opportunity to support this next phase of AI adoption. High-performing networks will be important to delivering the responsiveness, capacity and reliability that emerging AI experiences will require”, he added.
New AI habits create new network demands
Globally, a group of AI super users offers an early indication of how AI usage could evolve. These users make about 10% of all consumers today, and expected to increase to 30% by 2030. These users are generating more than 80% of all AI activity. Their behaviour points towards more frequent AI interactions across multiple devices and locations, with greater reliance on mobile connectivity.
For AI super users, response time has 2.5 times the influence on network satisfaction compared with average users, highlighting the significance of the overall experience as AI becomes more embedded in everyday activities.
Daily multimodal AI use is projected to grow from 10% today to 30 % globally by 2030. Sharing images, video and voice with AI services will increase upload traffic and create short bursts of demand. By 2030, daily multimodal AI use, multi-device AI use and the number of on-the-go locations, where AI is used are all expected to triple.
The research report titled An Agentic Life, also points to the growth of agentic AI, which can carry out tasks such as making bookings, comparing products or managing schedules on a user’s behalf. Across the markets surveyed, 38% of smartphone users expect to use agentic AI by 2030, compared with four percent today. Consumers expect these AI-enabled services to save around one hour of screen time each day.
As agentic AI becomes more prevalent, AI services could generate more background activity across applications and services. For networks, this could increase the importance of predictable performance, low latency and resilient connectivity.
Trust will also be important as AI becomes more capable of acting on behalf of users. More than half of daily AI users are strongly concerned about fake or hacked AI assistants. The research finds that communications service providers rank the most trusted among the ecosystem actors for verifying the identities of AI assistants.
The findings are based on responses from more than 43,000 smartphone users aged 15–69 across 27 markets, including India, Australia, Indonesia, Malaysia, the Philippines, Singapore and Thailand. Figures for 2030 describe anticipated usage, rather than current adoption.