Debunking The AI-Ready Data Myth

In coining the term AI-Ready Data more than 3 years ago … I was admittedly jargon-y.

In the proceeding years, I have realised that merely labeling a critical (enterprise) need with the genre du jour took away from the gravity & depth of what was required to make the idea of an intelligent enterprise a reality.

The push for data specifically engineered for artificial intelligence assumes that we are leaving the door open to many “next-gen labels”, without solving for the core problem.

Are we ready to start the din - ‘‘The World Needs Agent-Ready Data’? (sic)

Data, in its natural, diverse state, reflects the many permutations & combinations that make a functioning Enterprise - a complex web of systems, processes and humans. Precisely what makes that very data un-usable ‘as-is’.

Add to that the many rules, requirements, and constructs that must be reasonably enforced for an Enterprise to offer their products and services to society - and you have a challenge, to get data that is ready, for any use case.

Not just AI.

So restructuring data solely to suit the needs of AI will prove to be limiting.

It will also divert attention from solving for the most basic (yet important) of enterprise needs - data that is ready to be used for all wants.

Insisting on a single framework of “readiness” risks ignoring the variety of corporate, human, and systemic demands, therefore narrowing the usability of such data. Instead of expanding it.

So I am correcting my mistake.

The world does need data.

It needs to be accurate.

Accurate, not just for AI, but also for anything & everything else we may come up with in this fast & furious pursuit of being an inter-planetary super intelligent species!

   

PS - yes, i do believe in aliens…and no, i dont see dead people ;)

     

Image credit - Charles Minard's Flow Map of Napolean's Invasion