Announced at VMware Explore 2023Private AI Architecture Solution, allowing enterprises to introduce AI technology applications while ensuring data privacy and compliance with regulatory requirements. VMware's actual approach involves not only collaborating with external companies such as NVIDIA and Hugging Face, but also providing appropriate model frameworks through its internal AI Labs, enabling enterprises to quickly adopt and train applications, thereby creating AI application services that meet their needs.
However, how can VMware ensure that the model framework it provides is trustworthy or more in line with the actual needs of enterprises?
Chris Wolf, VMware vice president and head of AI Labs, said that the model frameworks provided by VMware will be checked by AI Labs to confirm that they will not cause privacy issues or have problems with the source of training data, and will not conflict with current laws. Only then will they be made available to enterprises to avoid causing greater problems.
However, Chris Wolf also emphasized that this does not mean that VMware will intervene and dictate the use of only specific model frameworks. In fact, it will continue to communicate with enterprises to understand their main usage requirements and determine which suitable model frameworks to provide in the Private AI architecture solution. Therefore, the main consideration is still based on the needs of enterprises, and it will not interfere with the application development of enterprises or developers.
On the other hand, in addition to collaborating with Meta, NVIDIA, and other industry players on large-scale natural language model frameworks, VMware will also continue to collaborate on open source model frameworks, or have large technology companies provide trusted model frameworks. VMware will even consider developing its own model frameworks for public cloud providers such as Azure and AWS.
With its Private AI architecture, VMware aims to help businesses that want to implement AI applications but lack the budget and capabilities to build appropriate model frameworks embrace AI application development opportunities more quickly. This will also reduce the risk of businesses using open source model frameworks that fail to meet their needs, potentially violating local regulations, or causing internal data privacy leaks.
In addition to ensuring that enterprises and developers can quickly and securely import artificial intelligence application resources using VMware services, VMware has also collaborated with Hugging Face to integrate the SafeCoder auxiliary tool with automatically generated artificial intelligence technology, allowing enterprises and developers to generate program code through natural language while also proactively checking for errors in the program code, thereby accelerating the efficiency of enterprise application service deployment in the cloud.
