The hype around AI shows no signs of slowing down and continues to attract attention from publishers, developers, anyone creating mobile applications for their business, and those with a keen interest. This fact is proved by the latest report, State of AI, which analyzed current trends in 2025, released by Sensor Tower, the leading source of mobile app insights. If you haven’t had time to look through the report, you can check out the most interesting trends from my perspective as a Product Director at ChatOn, one of the popular AI chatbots on Android and iOS with over 90M users, and built on various LLMs.
Generative AI apps continue to grow rapidly
If you don’t study the topic of AI, you should know that Generative AI apps can be divided into two main categories: AI Assistants and AI Content Generators. If we look at the overall trend in general, we’ll see that the number of global downloads has accelerated to nearly 1.7 billion in H1 2025. This number is impressive, given that the growth of this niche began only two years ago, with the introduction of ChatGPT by OpenAI.
Growth is particularly strong in Asia, which now leads the world in Generative AI downloads, outpacing both Europe and North America. However, North America still dominates in-app purchase revenue.
Consumers spent more than 15.6 billion hours in these apps in the first half of 2025, which means an average of around 50 sessions for every person on Earth.
Young men dominate among users of AI apps
The audience of AI Generative apps has a few visible trends. The largest demographic of users is young men aged 18-35. They account for more than 60% of users across all popular AI assistants, including ChatGPT, Google Gemini, Grok, Meta, etc. The female audience dominates in entertainment-focused AI apps like PolyBuzz and Character AI.
AI assistants have become a part of daily routine
Users regularly turn to their favorite AI personal assistants, and the frequency of sessions is increasing. In February 2025, users interacted with ChatGPT an average of 10 days per month; by June, that number had increased to 13 days per month. This trend is similar for all popular AI assistants. It shows that AI personal assistants have become a part of users’ daily routine. This frequency is similar to that in popular apps like X and Reddit, but it is lower than the frequency of using Google.
Users prefer to ask AI assistants questions about Lifestyle & Entertainment
Sensor Tower analyzed what topics are more interesting for users of AI personal assistants. The results show that users prefer to search for information related to Lifestyle and Entertainment. The percentage of these prompts increased from 22% in Q2 2024 to nearly 35% in Q2 2025. This may seem surprising, as AI tools were initially positioned primarily for education and work at the start of the hype. While the Work and Education category is also growing, its growth has been slower.
Among the other categories showing growth, I notice that Health & Wellness and Shopping are the most rapidly developing.
The importance of GEO is growing
As the percentage of users who search for information using AI chatbots is increasing, the impact of generative engine optimization is also growing. Brands should be visible to AI chatbots that form opinions about products and services based on criteria beyond those of traditional search engines like Google or Bing. The results of a search when a user asks the same question in Google and in ChatGPT can be extremely different. AI-generated answers to users’ questions are based on content from across the web, and this means that the visibility of your brand depends on the quality of your brand mentions. SEO, with its changing links and focus on keywords, will lose rankings.
Apps of all genres add AI features
The entire mobile ecosystem is feeling the impact of AI and is trying to adapt and implement AI features. This trend is seen across all categories, from calorie counters and nutrition apps to photo editing apps. The report shows that apps incorporating AI were downloaded 7.5 billion times in H1 2025, constituting roughly 10% of all downloads during that period. This also represents a significant 52% year-over-year increase in downloads for these AI-enhanced applications.
The popularity of open-source models is increasing
One more trend that I’d like to mention, despite its absence in the report, is that the processing of data directly on the user's device (primarily smartphones) has increased its popularity over cloud-based processing. This has been made possible by the emergence of specialized NPUs in mobile processors.
A key factor for this trend is the recent growth of open-source models, such as Meta's Llama 4s models. These models provide the flexibility to fine-tune solutions for specific needs and helps reduce costs for basic scenarios. Open-source models are ideal candidates for on-device deployment, enabling sophisticated AI features to run locally, thereby ensuring fast, privacy-oriented, and offline-capable operations.
I hope these highlighted AI trends in mobile app development will be helpful for you and inspire you to implement AI in your apps.
Top comments (0)