Recent trends in AI development emphasize on-device inference, leveraging small language models (SLMs). Google leads with its “Web AI” initiative, while DataSapien recently launched a mobile SDK for on-device AI applications, compatible with iOS, Android, and frameworks like React Native and Flutter.
The 20MB SDK features a low-code interface that grants developers access to numerous ML and SLMs. A notable aspect is its MeData vault for personal data storage, enhancing user privacy by keeping data on-device. DataSapien’s CEO, StJohn Deakins, describes its integration potential with external AI models alongside its own algorithms.
Use cases span retail and healthcare, with emerging functionalities resembling agentic AI, allowing proactive suggestions based on user behavior. DataSapien’s approach contrasts with Google’s Web AI, focusing on local compute to reduce costs while prioritizing user data security. As SLMs evolve, their potential in mobile applications becomes increasingly promising.
Source link
