Apr 2025

AI models need an Accountability Index

There are three (RIP to fourth) factions of people when it comes to AI.

1. AI is going to take Jobs for sure, entry-level folks are screwed. 😱 The top level will make more money. 🤑

2. AI is not going to take Jobs, it can help for sure- It will help everyone do their existing job better. 🚀

3. AI is going to create more Jobs. 💪

4. AI is going to end Jobs and everyone will just make art. ✨ (This bubble was the earliest to burst- AI became popular for the art, rap lyrics, and VanGogh imitations)

Depending on which faction they come from- they will look at this Google trends graph differently, make different inferences, and drive their point.

Then, there will be those who know this graph means nothing, and neither do the opinions of these factions without context.

The truth is - all of the first three conditions exist simultaneously.

It becomes a question of-

Why, where, when, and How.

Few Facts to consider-

1. AI has Lowered the barrier to entry- not in firms, but at "getting a job done".

I have negligible coding experience. I was able to create an Agent-based Model and run simulations from scratch using my insights and translating them in the MESA library assisted by DeepSeek r1.

I know friends from Bihar trying to set up Makhana business, and a lot of their current copy is coming from Chat GPT. They do not have the money to currently invest in Brand Professionals or Copywriters- but they have been able to get the job done.

How good is it? that's a different story. How good was my model? I won't know until I see the predictions playing out.

2. AI has given incentives to cut costs- but it remains unclear to most upper management- towards "What".

- To one firm, the creative minds are an asset. Their MVP. They look for automated paperwork.

- To another, their legal expertise is their stronghold, they look to automate their lead generation.

But this has always been the case with any solution-as-a-service offering no?

AI is trying to offer the same thing at lower costs- those who need it done at lower costs, because it is lower on their priority list- will take it up!

And then some Jhol will happen.

And then they will realise it is a matter of -

Where the Risk of AI automated work can be taken.

And where Human labour is necessitated, based on task complexity and criticality.

Each Job, each sector, and each Founder- needs to understand the ratio of these two aspects

Which task operates at what ratio of Task Complexity vs Cost of Failure.

This will help them understand an Index for allocating investments the right way- in Human and AI capital.

Why?

Mistakes by AI are still an "oops" by the system- a lot of offerings come with the caution, humans on the other-hand, can be held accountable.

It is, in fact, an accountability index.