Apple brings out an open source AI that runs directly on the device

[tasnimsanika00]

Apple has been working on artificial intelligence (AI) for a long time , but its efforts in this area have so far been rather silent. However, the company from Apple has finally decided to enter the AI ​​battlefield and it has done so with all its artillery. Those from Cupertino have just raised the curtain on eight large language models that are part of the OpenELM (Open-source Efficient Language Models) family and that, as their name indicates, are open source.

Google’s AI models are available for download from Hugging Face Hub , a platform for sharing AI code. And as you might expect from a company like Apple that has disruption deeply embedded in its DNA, it seems that the company led by Tim Cook wants to approach AI in a radically different way than rivals like OpenAI or Google.

OpenELM is an AI that is specially optimized for the powerful M chips lurking in the guts of benin whatsapp data Apple devices. And unlike models from other companies, models built on OpenELM provide shelter for substantially more material, including a complete guide to training them using publicly available data. The OpenELM family also includes pre-installed training protocols and configurations.

Apple's AI takes a stand on ethical and legal training
According to initial analysis, Apple's AI models are at least 2.36% more accurate than previously released large language models. They also boast that they require less training than other similar models. As open source models, Apple hopes this will help accelerate the research born out of OpenELM so that its AI can serve users with more reliable results.

The OpenELM family is the first step towards bringing AI to Apple devices. The models built into OpenELM are not necessarily at the mercy of a constant connection to cloud servers and are specifically designed to be executed directly on devices, which translates into faster response times and greater privacy for the user.

OpenELM only actually connects to the cloud if it needs to run more complex functions . In this way, Apple aims to prevent its AI from being fed with potentially copyrighted content. The Cupertino company's new AI models pride themselves on being trained following both legal and ethical precepts.

To avoid any potential controversy arising from the training of its AI models, Apple has also reportedly acquired licenses to be able to legally use content from major media groups such as Condé Nast on OpenELM . It is clear that Apple's approach to the AI ​​game is radically different from that of its rivals, and its huge user base will undoubtedly help it stand out from the crowd in this field of activity.