On Wednesday, Google (NASDAQ:GOOGL) unveiled new “open” artificial intelligence (AI) models dubbed Gemma.
The new initiative allows both individuals and businesses to create their own AI software using "open models" at no cost, following similar moves by Meta Platforms (NASDAQ:META) and other tech companies.
Google is also sharing crucial technical information, including model weights, with the public.
With Gemma, the Alphabet-owned company aims to draw software developers towards Google’s platform and boost engagement with its now-profitable cloud division. Moreover, these models have been fine-tuned for optimal performance on Google Cloud, where new users are incentivized with $300 in credits.
“Gemma is built for the open community of developers and researchers powering AI innovation. You can start working with Gemma today using free access in Kaggle, a free tier for Colab notebooks, and $300 in credits for first-time Google Cloud users,” Google said in the official announcement.
However, the tech behemoth has decided not to fully open-source the Gemma models, retaining some control over their usage and terms. This decision comes amidst ongoing debates about the potential for misuse of open-source AI, despite its benefits in democratizing access to technology.
In contrast to Gemma, Google did not make its larger, flagship Gemini models open. The company highlighted that the Gemma models are developed with either two billion or seven billion parameters, which represent the diverse values an algorithm uses to produce results.
In a related development, AI chip designer Nvidia (NASDAQ:NVDA) said it has collaborated with Google to ensure the seamless operation of Gemma models on its hardware.
Moreover, the company also said it is in the process of developing chatbot software designed to operate AI models on Windows PCs and plans to make this software compatible with Gemma.