Gemma AI promises to make chatbots safer for users

Google has introduced Gemma, an open-source AI model that will let users to construct their artificial intelligence chatbots and products using the same technology as Google Gemini (previously known as Bard and Duet AI).

Gemma is a collection of open-source models curated using the same technology and research as Gemini, created by the Google DeepMind team. Along with the new open-source strategy, Google has released a 'Responsible Generative AI Toolkit' to help developers get started and experiment with Gemini, according to an official blog post.

The open-source model is available in two variations: Gemma 2B and Gemma 7B, both of which have been pre-trained to filter out sensitive or confidential information. Both versions of the model have also been evaluated with reinforcement learning from human input, which dramatically reduces the likelihood of any Gemma-based chatbots spitting out dangerous information.

While it's easy to dismiss Gemma as just another model that can generate chatbots (and you wouldn't be wrong), it's worth noting that the business appears to have really built Gemma to "[make] AI useful for everyone," as stated in the release. It appears that Google's current model is intended to encourage more responsible usage of artificial intelligence. 

Gemma's release comes just after OpenAI released the stunning video generator Sora, and while we'll have to wait and see what developers can create with Gemma, it's encouraging to see Google try to approach artificial intelligence responsibly. OpenAI has a history of releasing features and products, then cleaning up the damage and installing protections afterward (in the spirit of Mark Zuckerberg's 'Move fast and break things' adage).

Another intriguing aspect of Gemma is that it is meant to run on local hardware (a single CPU or GPU, but Google Cloud is still an option), implying that something as basic as a laptop may be used to develop the next popular AI personality.