How to better measure changes in standard of living in a world accelerated by AI and technological progress ?

(This article was a contribution to the work of the Live With AI think tank I'm a board member of, and was also published in Les Echos, the French leading financial news organization)

Why GDP/capita and purchasing power are insufficient tools in a world of AI and fast technological progress ?

The Gross Domestic Product (GDP) is a metric refined by John Maynard Keynes in the 1940s to measure the economic output of a country. It's used as well as a proxy for "economic power", while GDP/capita is widely construed as a measure of standard of living, and even sometimes welfare.

We've long known that GDP is broken, the list of its ills is a long one : the more one country depletes finite natural resources, the bigger the GDP. The more prisons one country builds, the more drugs and paid-sex it consumes, the larger its GDP. The US spends more per capita on health care than France, yet an American on average has a poorer health than a Frenchman. Are we confusing the end with the means ?

Also, GDP doesn't care about how much pollution is generated in the process. And conversely, some activities beneficial to the economy are not taken into account. Like a household mum patiently having her kids do their homework, or someone caring for an elderly relative. It led the economist Paul Samuelson to joke that GDP falls when a man marries his maid.

GDP/capita doesn't say anything about income distribution, and is imperfectly correlated with purchasing power. Without knowing how many hours got worked, it doesn't speak of variations in productivity between countries. And it doesn't account for how physically and mentally demanding any work actually is.

Technological progress and AI make things worse

One classical example is with the recorded music industry whose revenues have plummeted since its peak at the end of the 20th century and the advent of the internet. Music sells much less than before, hence contributes much less than before to GDP. But more people than ever can access more music than ever, more easily than ever, everywhere, all the time. What on Earth is going on here you might ask ?

 Even weirder for goods or services which became free altogether ! Calling a relative at the other side of the world used to be prohibitive, setting you back dozens of dollars. Now it's free. We get more comfort, yet that calling activity has disappeared from our calculation of standard of living when measured by GDP/pax or purchasing power !

Yet another confounding example is to be found with our digital devices and computers. True, the best of them today don't cost much less than before. But they are tremendously more powerful, have better memory and computing power, higher resolution, and gives us access to plenty of free apps and free entertainment.

Same with cars. Looking at the inflation-adjusted price of a car in the US for the past 50 years, one can see that it has basically remained the same ! But in the meantime quality of cars has gone up dramatically : they pollute less, are more fuel efficient, more comfortable, and above all much safer.

The issue that GDP is facing here is that gains in performance and quality are not well accounted for, if at all. Some advances in GDP measurements were made in the past couple of decades, like for instance with hedonic prices, but it's still far from perfect.

How meaningful are concepts of GDP/capita and purchasing power while a world of abundance spearheaded by AI is near ?

What happens in a world where with little human help, smart and dexterous robots and programs design, make and maintain other robots and transportation systems ? A world where such robots recycle matter endlessly or go get some more in our Asteroid Belt if needed ? A world where we've learned how to harvest the quasi-unlimited energy of the sun and the atom ? 

Based on the pace of progress in robotics, 3D printing, AI, nanotechnology, biotechnology, vertical farms and aeroponics, energy and space technologies, among others, we should not rule out a world of abundance in our century, at the very least when it comes to the goods and services required to live decently today. That would be a total game changer for humankind. We, or our children may live to see its advent ! 

In such a world of abundance, we would have access to more than we can consume or even spoil, as our ability and time to consume and spoil will likely be inferior to the ability of smart robots to build, produce, fix, repair and recycle. This is assuming human population will stagnate (as seems likely by 2050 according to demographers) while the robotic population can scale to whatever extent is needed to maintain the abundance, given the sheer quantity of energy and matter available in our Solar System. This also assuming AI will reach its potential as the new defining general purpose technology of our era, infusing tech and the economy at all levels.

So, were such a world of abundance to materialize, were a world with nearly unlimited robot labor, matter and energy to come, wouldn't the very notions of price, purchasing power and GDP likely be disrupted ?

In an ever more digitized and AI-enabled world, we need to reconsider how to best keep track of the evolution of standard of living. As GDP/capita and even purchasing power accumulate flaws, why not use the new concept of "access power" to measure for someone the evolution over time of the quality and array of all goods and services available for a given income, including free services ?

We can access more and more things for less and less money, sometimes even for free. All things considered, for many of us it feels like our access power keeps increasing, even if it's not globally shared yet unfortunately. This concept definitely calls for more clarity and research but it should make for a more optimistic conversation around AI !

For sure, such an access power concept is bound to be more subjective than normative, and probably hard to use to compare countries. But if it can do a better job at tracking evolution of standard of living over time within a country than what we have, that's already a start.

Maybe the access power has to be defined for different, subjective categories. At the end of the day, maybe it's up to any organization to launch its own access power gauge based on subjective criteria to be defined clearly ? Then up to the rest of us to find it useful or not, and the best ones will stick. 

Maybe for instance we could think of an access power index for information that will take into account and track over time for a given country the average internet speed, price and coverage, levels of censorship, among other things.

One could easily come up with such subjective indexes for many fields of the human experience. As long as they're useful and telling enough of the changes in quality we're witnessing across time and space, that's a step in the right direction.

______________

About the author : I'm a French entrepreneur based out of Singapore, and the cofounder of digital marketing agencies KRDS et OhMyBot, with offices in 6 countries. I'm passionate about new technologies and AI and their impact on humankind. Find my other articles and thoughts on my webpage and on twitter.