Oracle Seeks AI Automation Patent As Cloud Demand Grows

Oracle may want to hand off IT work to AI. 

The company is seeking to patent a system for AI-based software “patch management.” As the title implies, Oracle’s patent details a system that uses a predictive AI model to optimally target appropriate software patches.

“Deciding which patches to install or which patches take priority over others may be difficult when there are thousands of patches available at any given time,” Oracle said in the filing. “Manually identifying and applying security patches is time-consuming, labor-intensive, and error-prone.” 

First, this system collects metadata related to “target assets,” such as data center servers, computers, or other network devices, within a certain IP address range, and analyzes them to find potential security vulnerabilities and the corresponding patches that would fix them. 

Oracle then employs a prediction model to decipher which patches apply to the vulnerabilities on these devices, filtering out irrelevant or unnecessary patches. The refined list is then presented to the user (likely someone on an IT or security team) via a user interface, and the user can approve or reject each one to refine the list even further. 

The system finally automatically creates an “electronic change request” and installs the list of approved patches. 

While Microsoft, Google, and OpenAI have dominated the headlines over the past year for their eye-popping AI investments and research, Oracle has sought to keep up, too. The company announced plans earlier this year to boost generative AI capabilities within its cloud infrastructure, adding to the traditional AI it’s already employing. Patents like these may reveal the inner workings of those systems. 

One of the biggest beneficiaries of the increased appetite for AI and computing power has been cloud service providers. While this is Oracle’s bread and butter, the company is constantly battling for market share against AWS, Google Cloud, and Azure. Oracle only commands around 2% of the market, according to CRN. 

But the company is seemingly making a comeback after its fourth-quarter results last week revealed a backlog of to-be-fulfilled contracts that totaled $98 billion, primarily driven by cloud demand to train AI models. Its shares jumped following the report, leading to the stock’s best day since late 2021. 

The company also announced a database partnership with Google Cloud, and a deal with Microsoft and OpenAI to help fill computing needs for its AI models. OpenAI said in a post on X that its relationship with Microsoft remains “unchanged,” but working with Oracle “enables OpenAI to use the Azure AI platform on OCI infrastructure for inference and other needs.”