Microsoft is gearing up to intensify its integration of AI-powered functionalities into Windows 11, with the imminent release of Windows Copilot this fall. However, their ambitions extend beyond this milestone. According to information from Windows Central, Microsoft is currently in the early stages of experimentation with novel features for its native Windows applications, including Photos, Snipping Tool, and even Paint, all falling under the expansive domain of “AI.”

Reports suggest that applications like Photos, Camera, and Snipping Tool, primarily employed for handling images and screenshots, may soon incorporate optical character recognition (OCR) capabilities. These enhancements would empower users to effortlessly extract and paste text from images directly into word processors and text editors. Additionally, the Photos app could acquire the ability to identify individuals and objects in images, simplifying the process of separating them from their backgrounds.

Conversely, the iconic MS Paint application could undergo a transformation with the infusion of generative AI features. These features might enable Paint to generate images based on textual prompts, akin to the capabilities currently found in advanced image editors such as Adobe Photoshop. It’s worth noting that Microsoft’s Bing Image Creator already utilizes a DALL-E-based model to craft AI-generated images.

The implementation of some of these features may hinge on the presence of a neural processing unit (NPU) integrated into the user’s PC processor. Although NPUs have been part of Qualcomm’s Arm processors for some time, traditional x86 PCs powered by AMD and Intel processors have not included them until recently. AMD’s recent 7040-series chips and Intel’s upcoming “Meteor Lake” refresh are set to incorporate NPUs.

At present, Windows 11 boasts only a handful of NPU-exclusive features, predominantly focused on image enhancement and background replacement during video calls. Local NPUs enable more AI-accelerated tasks to be processed directly on the user’s computer, reducing reliance on cloud resources. This addresses privacy concerns and mitigates model-training issues associated with AI-powered products.

Many of these forthcoming features appear to align with the category of unobtrusive, broadly beneficial functionalities that were once grouped under the “machine learning” umbrella in earlier times. It’s important to note that Apple has already integrated similar character-recognition capabilities into its Photos applications for macOS and iOS, harnessing the power of their “neural engine” embedded in A- and M-series processors over the years.