A few days ago Google announced the arrival of some new large language models (LLMs) belonging to the family Google Gemma. The peculiarity of these tools is that they are very similar to Google Gemini despite being open source (and therefore free) and much lighter than their “big brother”.
Big G’s goal is to give developers an efficient system to use for their projects and, in this specific case, oriented towards creating AI chatbots.
What we know about Google Gemma
According to information shared by Big G, these models were created to “build artificial intelligence responsibly“, also releasing a guide called Responsible Generative AI Toolkitto help developers use these technologies in the best possible way.
There are currently two versions of Google Gem available: Gemma 2B e Gemma 7Band both can be performed without any difficulty even on notebooks and without the need for special resources.
Despite this, however, Big G also guarantees high performancecapable of satisfying any need of developers, who can “hold” a tool in their hands powerful, efficient it’s completely free.
Google also confirmed that the training took place only on data carefully selected by the research team and therefore considered safe. In addition to this Big G also guarantees a certain reliability of informationverified via a feedback system coming from real users who have validated the content.
Finally, the information used was private of all personal data of the users who participated in the model training so as to guarantee them maximum confidentiality.
In the coming months, the Mountain View giant confirms, they will be released new variations of Gemma so that their use can be oriented towards different types of applications, different from AI chatbot.
What future for Google’s AI
Following a strategy of this kind, Google (and many other giants in the sector) are increasingly reorienting their work towards these open source modelsdeveloped specifically to offer companies and developers tools that are reliable and to use without subscription.
On the other hand, with a choice of this kind, it is clear that Google aims to “involve” an ever-increasing number of users in the choice of its products so as to conquer an increasingly important share of the market, exactly as happened with Android.
According to experts, the future ofartificial intelligence it is open source and, probably, this prediction is not entirely wrong, especially because technology giants like Big G are not exclusively interested in earnings, but consider it equally important user data to be used for training your own linguistic models which represent a real asset of inestimable value.
To know more: Artificial Intelligence: what it is and what it can do for us