Obviously a rhetorical question, this is an adaptation of the phrase “robots will do the work, but not buy the cars” when the big shop-floor automation transition was happening in the US in the 1970s.
So while AI do not take lunch breaks, that is also because they do not consume anything. Not food, not anything else. Neither in urban areas, not in rural areas. Not in happy times, not in distress. They do not buy financial products, consumer products, houses, aataa + daal, and certainly do not indulge in luxury spends.
Now AI is expected to drive huge productivity gains - anywhere from 10% to eliminate certain roles altogether. We already see the biggest gains happening coding (via co-pilots), where the adoption rate is around 25% per Goldman Sachs. And data points in that direction in India too - where the IT sector in India was severely resource constrained - it has flipped a magnificently surplus market - all within a span of 3 years! So much so, the painfully atrocious persistent attrition rate of ~20% in the IT services companies is down to ~13%.
So what happens when the 25% penetration moves to 50% in the IT. And the average across all industries move to 25%?
With this backdrop, it was indeed then a timely reminder by CEA V Anantha Nageswaran yesterday, during the Economic Survey’s release, where he did prod the industry to mull over if AI will indeed cripple our consumption story. To spur our collective imagination, he produced two charts which basically show the decline of proportion of labour in total income (from 72% in FY18 to 54% in FY23). In other words, the income is shifting towards capital providers vs labour.
Which in itself would be worrying.
The Sun will continue to shine
Consider this: during the 1930’s Great Depression, ~the US unemployment was 20% when the legendary economist John Keynes declared we have a new “disease”, that is, “technological unemployment”. His concern was the economy was not finding many productive uses of the labour that was freed up from switching telephone lines (the most common for American women), not tilling the farmland, not pulling the cart, etc.
The tech progress continued. Productivity growth in the US between 1947 to 1960 was 3% - not a small number by any stretch.
But then new jobs come in too, and as MIT’s David Autor has shown - over 60% of the jobs in today did not exist before 1940s.
My view is the future jobs will be very different. And the folks in the middle would be hit hard. No less. And hence they will have to adapt quickly. But there are enough new skills that are up for picking - and frankly not enough folks have a solid handle on these.
So yes, pick new skills.
And for anyone who’s getting to college now, my advise is two-fold:
A) Learn 3 core skills:
- technical,
- creative,
- critical thinking
These will be your lifelong anchors. You can expand in many directions using these as a springboard. Besides these are not by way of formal training. You can pick them up on the side, or by way of certifications, etc.
B) Work on developing 3 soft skills:
- expressing ideas,
- building relationships,
- remaining ever curious
And in wrapping up, here's some data to bring cheer.
The unemployment rate in the US today is ~4% (on 3x the population compared wtih the 1930s). And by the way, 85% of all employment opportunities created since 1940s have been technologically-driven creations. That’s per MIT.
So that should, hopefully, allay all concerns about “technological unemployment”.
#ArtificialIntelligence #FutureOfWork #ProductivityGains #TechnologicalUnemployment #SkillsDevelopment #AIImpact #EconomicTrends #Innovation #JobMarket
