Generative AI, AI Tech's Impact On Jobs Will be Significant and Varied

Some years ago we hired Raju as an office boy at Tonse. He was a migrant worker from a small town near Tumakuru and had never worked in an office before. His father's farms were not productive due to the lack of water and there were few jobs in his village. He came to Bengaluru by bus one day to find a job and eventually set up a photo studio with his meager savings. Someone cheated him and promptly he lost all his money and ran away from the landlord. When we hired him, he was glad he got a job.

He quickly learned the ropes of working in an office - making coffee, ensuring the office was clean and neat, getting courier and other little things done, and soon became the most important handyman you couldn't do without. If Raju was on leave, everybody groaned and moaned that day.

Raju grew as we grew and he soon had a motorcycle, and even a junior to help him with his daily chores at work. In six years he had upskilled significantly, spoke some English, got a car driving license and could write his CV, became computer proficient, and finally moved to Amazon in their packaging section.

That was five years ago. I met him recently and he is now a confident young man, sends his kids to a good school, and is happy and busy. I asked him what he was up to. He said he now makes Rs 25k a month delivering milk in the morning for a premium milk brand for 4 hours in the morning and the rest of the day, he makes another similar sum delivering organic pharmaceutical products from his home for orders that arrive on his mobile app. All his work is mostly within a 4 sq km area around his house - has plenty of free time, spends time with family, and attends Friday satsangs in the local temple.

All good so far.

When I asked him if he could scale further by bringing in junior delivery staff under him and making a small team so he could serve more customers, he said - the biggest issue is a lack of dependability. If a delivery runner suddenly takes a day off, it is hell for him as there is nothing he can do at the last minute to find a replacement. So he has decided that he will always operate with a size of one.

To me, this was an ugly reminder of my Consulting days, while taking on new work was a challenge unless you brought in qualified consultants. They cost a ton of money and to keep the team going you need to continuously take on new projects. Experienced consultants tend to leave seeking greener pastures, and replacing qualified specialists is next to impossible. The treadmill grinds on - wearing you out in the revenue, execution, and reputation triad that eventually burns you out.

How soon had the model trickled down from hi-tech consulting to milk delivery? I thought.

As Dunzo, Zomato, and Swiggy have realized by now, the volume at the bottom of the pyramid in terms of people available to take on low-skill tasks such as food delivery (industry parlance is 'runners') does not necessarily offset the inherent inefficiencies in skillsets, and float (high turnover in runner count).

They simply hire more or 'bench' (as we call it in the software industry) to factor in absenteeism or surge business. This does not make the model any more efficient but bloats the cost further.

The point is, the service industry rises to the level of human inefficiency and then stays there at best or collapses at worst. Raju can control only his availability and commitment and will be good to take on only so much commitment - and no further. By the way, this is well-known and nothing new.

In my talks with another semi-skilled workforce, such as the gardener, car drivers, plumbers, carpenters, and other such workers in the communities I found the same issues. They have plenty of work to take on but they cannot delegate and cannot take on more work. For this set of people, finding work is not an issue anymore but finding the time and energy to do it is. Why? Because they cannot delegate, subcontract, or outsource further downstream.

Generative AI / AI Tech Will Subsume Jobs in Skilled Sector

The interesting thing about what generative AI and soon AI, in general, will do is impact from the top down. Look at the type of impact AI technologies have been able to do in the last 90 days to get a picture of where we are heading

Architects and Interior Decorators

• 360degree fly-through of a house design -can match a drone view or better - with offline integration of 2D still images

Impacting - architects, interior designers, visualizers, home and office design agencies

• chatGPT and LLMs running on low-level hardware like Google Pixel 6 and 4GB RAM Raspberry Pi4

Impacting -expect all kinds of hardware, including Edge devices, laptops, and GPU to carry pre-built offline GPT/AI for personal use. Should not be surprised if Microsoft Surface, tab /laptop starts the earliest

Coaching, Marketing, Research, Accounting

These functions are largely going to get automated at the least base level if not end to end. There is an insane amount of applications that are being built over gen AI and most will come free or very low fee. Already finance-specific version is being made available by Bloomberg and such specialized domain-specific applications will be in plenty. Some estimates are that these could retail at as low as $0.50 cents an hour.

Call center jobs

Deep fake voice AI applications scraping youtube videos for voice signatures to be carried into personal phone calls are so effective that people are being ripped off in early fraud cases just being reported in the last 3-4 weeks. Why would thousands of call center agents be required if a single laptop with a subscription to an AI voice engine can do the same - for a few cents per profile?

One shudders to think of what would happen to a 10k seat call center operation in Philippines or India at this rate. The software will be far more accurate, personalized to the voice the customer would like to hear, play the exact text required from the same script the human agent used, and do it for a fraction of the cost (and translate to multiple languages and time zones with a flip of the switch).

It is impossible to keep up with the new releases on generative AI wrappers that are becoming tools for artists and creators to interface - the design outputs being generated by the tool itself - many times in a few minutes and mostly for free.

Look at the level of jobs at risk due to automation and new technologies taking over even creative functions. Automation is therefore going to be lethal in several mid-level jobs and not just boring repetitive jobs. And this could happen within the next 18 months.

Lower down the ladder, semi-skilled jobs such as plumbing, and diary delivery (unless you are talking about drone delivery which is at least a decade away in India) are likely to hold on for some time.

Unfortunately corporations that have these positions will book higher profits as AI partially or fully replaces these jobs. Governments will stay mum and these qualified folks may be left in the lurches not knowing which way to go.

The public debate on the validity, speed and mass implementation of Gen AI/ AI tech has not even started. Yet @Microsoft, and @Google are expected to go into full implementation to have these capabilities into the hands of the masses at the earliest. There must be a better understanding of how these things work and to what extent they should be allowed to be implemented - else much like social media technologies that were given a free hand in the last decade, we might end up putting the cart before the horse and wake up a trifle too late?

About the Author:?Sridhar Pai Tonse?writes about life, tech, markets, and startups. He is an expert on Strategy and Lead Generation for tech startups. Follow him on?https://youtube.com/@tonsepai ?and visit?https://tonsepai.com . For more?https://tonsetelecom.com . Twitter:?Sridhar Pai Tonse — Marketer, storyteller, coach