Saturday, January 27, 2024

Is AI a Platform, an App, a Use Case, a Function or a Capability?

Is artificial intelligence a platform or an app? Is it a use case, a function, a capability or an ecosystem? Might it be any and all of those things, depending on the context? 


Though it might be most common to think about artificial intelligence as a platform rather than an app, use case or function, AI might be a platform in one sense; represent use cases in other senses; or functions or even discrete apps. 


Think about the internet, something most observers might agree is a general purpose technology and therefore a “platform.” Recall that a platform is “hardware or software that other hardware and software can run upon.” 


In other words, the platform is the infrastructure upon which apps are built and executed or run. But we commonly will encounter use cases where AI “seems” to be an app, a use case or function. 


Used on a smartphone to support camera features, AI might act like a function, allowing us to manipulate images. 


In other cases, such as speech recognition, we might consider a branded capability--such as Siri or Alexa--to be akin to an “app.” 


And sometimes--such as with automated vehicles--the use of AI might be so intertwined with a product that it is a use case. 


Online marketplaces (Amazon, Etsy) are platforms in a business sense.  Social media platforms (Facebook, LinkedIn) are platforms in that same sense. 


Perspective

Description

Example

Platform

Provides foundational tools and capabilities for building various AI applications. Think of it as the "Lego set" for AI developers.

TensorFlow, PyTorch, Azure Machine Learning

Use Case

Specific problem or task where AI is applied to deliver a solution. Focuses on the "what" and "why" of using AI.

Medical diagnosis, fraud detection, self-driving cars

Function

Individual AI capability or algorithm that can be combined with others to build applications. Think of them as building blocks with specific functionalities.

Natural language processing, image recognition, anomaly detection

App

Software program that integrates AI functions to deliver a specific service or user experience. The finished product, ready for end users.

Siri, Alexa, facial recognition software, personalized recommendation engines


And that might be a key to our general sense that sometimes AI is a platform; sometimes a use case; sometimes a function; sometimes an app. 


Think about the computing GPT. Though we might sometimes talk about “using computers,” as in “using a smartphone or a PC or tablet,” usually we are really talking about “using an app,” taking advantage of a function or invoking a use case. 


We might open a social media app to pass the time or find out what’s going on. In that instance we might consider that we are “using an app.” 


At other times we need information, such as what is next on a calendar, or how to get someplace. That might be more of a function. Yet at other times we might want to buy a plane ticket, in which case we might consider that computing instance to be a use case. 


So AI might “seem” a platform, or an app, or a use case or function, depending on the context where it is invoked or consumed or used.


No comments: