Majority of the mundane tasks done by consultants and tellers have been outsourced to the customers themselves. Let us take an example of replacing your washing machine. You have to start with the reviews of the various companies in the industry, their customer support, plethora of models, price, features and ...... Then you finalize by comparing various by discussing with friends and other customers. Then you decide where to buy by visiting various websites. By the time you make the decision you will not know how much time you have spent in buying the washing machine.
The above example is just a tip of the iceberg. The iceberg of activities outsourced to the customers themselves. Now, I believe we will have AI will debut in helping customers buy products and services. In fact it should not help, it should decide on the customers behalf. Not again curated links to various web pages. Why? Well that is as simple an AI can get and the amount of data available on buying habits and preferences are abundantly collected by search engines and social websites and blogs. Just imagine our preferences which are now embedded in search engines, apps, social networks, instagrams and the like..
The time we wake up, what time you leave your house, your vehicle, the breakfast we have, distance between your work spot and your home and everything in between up to the book we have been reading before we sleep, how many pages we average everyday and may be map our dreams based on the week day and the way our day was and also alter the dreams accordingly for a pleasant sleep.........
In the context of AI as the first step, the data sources of the following are crucial. In AI, the belief system is the most important link. The belief of the user who generates data in the mining company. The belief of the mining company in the user data generated regarding the authenticity of it. The belief is mutual and should not be compromised at any point by both the agencies. Coming back to the data sources
a. The personal data
b. Generic data available on the public web/sources(Like the mileage of a car)
The simplest AI would be to combine the personal data(to be mined from conversations, posts, comments, search history, buying pattern etc. This is to begin AI. Later as the user plugs in to the actual AI, data collection will be very specific and more to the point avoiding "assumptions" on the part of AI) and the generic data from public domain(information culled from official websites). Here again data integrity should be of the highest level.
The whole experience of buying should be hassle proof and painless. I start by telling the AI of my intent in buying a washing machine. Processing the personal and generic data majority of the decisions are be made and only when contradicting data is reached, user should be prompted. Here again, the prompts should not be beyond two, after which, the experience gets frustrating with the irritating prompts.
We definitely are now in position for the first step in AI. But, the innumerable protocols(open and sucking proprietary), the idiotic app ecosystem, wherein the data gets boxed, the plethora of operating systems and data integrity are the biggest issues combined with Internet connectivity across the world to the last mile. It is time all the data keepers start collaborating by providing API's to supply and share data among themselves for the purpose of AI. The race should be an open and fair one. The race should only benefit the customer. The race to be won by the best AI, not on the amount of data available with each other but by its capability to take decisions on behalf of the customer.
All this brings us to the biggest question regarding the biggest company with the largest amount of data, Google/Alphabet. How will google/alphabet make money out of AI. Will it charge the customers like maintaining a personal assistant. Only few afford personal assistants. You have to push your personal assistant to the user and the way to do it is not charging the customer. How will google/alphabet make money here, when the user is not provided any web page to click? Will google/alphabet tweak the AI to take decisions based on the amount payed by the manufacturer, similar to improving page rankings. That is a failure when it comes to AI. It has to be fair. If it is not fair, people will ditch it for alternates. Now alternates would have learnt their lesson. How will google\alphabet make money? HOW?
Well, Google/Alphabet, I am looking at you. Bring the AI on. Bring it on. Bring on the beta. Let us start. From us being a product take us to what we are, Customers. We are the customers and we are not the product. This is the basic change you have got to do for succeeding in this AI endeavour. If you don't treat us as customers, your AI engine will be flawed. Keep us in the God's place and make AI our slave. Will you?
The above example is just a tip of the iceberg. The iceberg of activities outsourced to the customers themselves. Now, I believe we will have AI will debut in helping customers buy products and services. In fact it should not help, it should decide on the customers behalf. Not again curated links to various web pages. Why? Well that is as simple an AI can get and the amount of data available on buying habits and preferences are abundantly collected by search engines and social websites and blogs. Just imagine our preferences which are now embedded in search engines, apps, social networks, instagrams and the like..
The time we wake up, what time you leave your house, your vehicle, the breakfast we have, distance between your work spot and your home and everything in between up to the book we have been reading before we sleep, how many pages we average everyday and may be map our dreams based on the week day and the way our day was and also alter the dreams accordingly for a pleasant sleep.........
In the context of AI as the first step, the data sources of the following are crucial. In AI, the belief system is the most important link. The belief of the user who generates data in the mining company. The belief of the mining company in the user data generated regarding the authenticity of it. The belief is mutual and should not be compromised at any point by both the agencies. Coming back to the data sources
a. The personal data
b. Generic data available on the public web/sources(Like the mileage of a car)
The simplest AI would be to combine the personal data(to be mined from conversations, posts, comments, search history, buying pattern etc. This is to begin AI. Later as the user plugs in to the actual AI, data collection will be very specific and more to the point avoiding "assumptions" on the part of AI) and the generic data from public domain(information culled from official websites). Here again data integrity should be of the highest level.
The whole experience of buying should be hassle proof and painless. I start by telling the AI of my intent in buying a washing machine. Processing the personal and generic data majority of the decisions are be made and only when contradicting data is reached, user should be prompted. Here again, the prompts should not be beyond two, after which, the experience gets frustrating with the irritating prompts.
We definitely are now in position for the first step in AI. But, the innumerable protocols(open and sucking proprietary), the idiotic app ecosystem, wherein the data gets boxed, the plethora of operating systems and data integrity are the biggest issues combined with Internet connectivity across the world to the last mile. It is time all the data keepers start collaborating by providing API's to supply and share data among themselves for the purpose of AI. The race should be an open and fair one. The race should only benefit the customer. The race to be won by the best AI, not on the amount of data available with each other but by its capability to take decisions on behalf of the customer.
All this brings us to the biggest question regarding the biggest company with the largest amount of data, Google/Alphabet. How will google/alphabet make money out of AI. Will it charge the customers like maintaining a personal assistant. Only few afford personal assistants. You have to push your personal assistant to the user and the way to do it is not charging the customer. How will google/alphabet make money here, when the user is not provided any web page to click? Will google/alphabet tweak the AI to take decisions based on the amount payed by the manufacturer, similar to improving page rankings. That is a failure when it comes to AI. It has to be fair. If it is not fair, people will ditch it for alternates. Now alternates would have learnt their lesson. How will google\alphabet make money? HOW?
Well, Google/Alphabet, I am looking at you. Bring the AI on. Bring it on. Bring on the beta. Let us start. From us being a product take us to what we are, Customers. We are the customers and we are not the product. This is the basic change you have got to do for succeeding in this AI endeavour. If you don't treat us as customers, your AI engine will be flawed. Keep us in the God's place and make AI our slave. Will you?
No comments:
Post a Comment
Nobody can deter me away from "free as in freedom" concept seeded by Sri RMS. See to it that u dont make fun of my belief. If u think otherwise, no need to comment.