AI Development Finally Accessible to All
In a positive development, Google’s Gemini Nano, an
advanced on-device AI model, has kicked off for Android Developers. Now using
the AI Edge SDK on the Pixel 9 series, developers are now able to utilize
Gemini Nano without an internet connection which is a huge milestone in
on-device AI.
Expansion of previously limited access
As per the findings, Gemini Nano was made publicly
available through the Pixel 8 Pro which came out in late 2023 to a few Google
applications exclusively. Further, it was gradually made available for the
Galaxy S24 series and recently Xiaomi devices. However, Google took an
additional step in making it available to the general Android development
community which perfectly underlines the leaning in making on-device AI more
mainstream.
What Developers Need to Know
As of now, developers can use text-to-text prompts
which allow for such tasks as:
– Rewording Texts
– Auto Responses
– Editing Texts
– Information Extraction
Each of these text-based tasks provide the developers
with a valuable context into what the architecture of the production device
offers, although mass production utilization is not allowed as the system is in
deeming stages.
Future outlook – Multimodal and Broad
Integration
Support for certain import modalities, such as images,
was recently announced by Google. This could enable AI developers to move into
more types of media. Expansion may also be possible to all current and future
models by Xiaomi, Motorola, Realme and other device makers, making Gemini Nano
available on more Android devices.
Beneath the Surface of Getting Started
with AI Core and Gemini Nano
Gamers that want to test out Gemini Nano for
themselves will be able to join Android’s AIcore (Beta) system via Google’s
Play Store. This new entry might stimulate the creation of a number of new Ai
powered applications on the Play Store thereby securing Android’s stake in the
AI apps market.
