<- home

GitHub
Thumbnail

AI & Wealth Inequality

Concise thoughts on the relationship between AI and income inequality.

Artificial Intelligence will help elevate people out of poverty

Technology has significantly improved human living standards. Global GDP barely changed in the roughly 1800 years between 1 AD and the beginning of the Industrial Revolution, but witnessed a parabolic 57,000% increase over the subsequent two centuries from 1800-2015.

Although there were undoubtedly confounding factors, we can be fairly certain that echnological advancements have greatly enhanced human living standards, as seen in the increase in the poverty threshold and the remarkable growth of GDP.

Applications of AI in improving human development index (HDI)

On the surface, it seems like AI is likely to aid humanity in continuing on this trajectory. The following are likely applications of AI in the near future;

  • In healthcare, AI will be instrumental in aiding discoveries and revolutionizing disease prediction source, source.
  • Applications of AI technology in agriculture can significantly improve our producion source.
  • In education, AI can improve access and efficacy source.

The potential of AI to exacerbate income inequality

The introduction of new technologies into societies brings about the real risk of loss of employment, as the technology warps and evolves the requirements of the society. Given the nature of Artificial Intelligence, it seems highly likely for AI to warp the society within which we live more than the advent of any other technology ever intvented. AI that can reason about the world we live in and make complex decisions in dynamic environments would be able to automate the work of many millions of white collar workers.

In the near term this includes the jobs of transcribers, translators, data entry and analysis workers, and generally speaking jobs that are predominantly based on reading and writing. In the medium to long term, we might expect this to spread to domains that involve short term reasoning and decision making. Jobs that involve logistics, planning and organising might best be done by machine intelligence that can juggle all of the many complexities involved simulataneously, whilst exploring and analysing many millions of possible future trajectories at once to find the most suitable course of action.

At the same time, the newfound increases in productivity and resources will require increasingly more orchestration and maintainance. In the same way that the advent of work done by machines increased employment by creating news jobs (jobs such as the maintainence of machines), so too will AI create new jobs.

It appears however, that as intelligence becomes cheap, the nature of the job market will change, with the newly required jobs becoming increasingly more technically demanding source, source, source. Jobs like data scientist, machine learning engineer, system architect and capital allocator will become more and more common.

Is inequality bad even if everyone is getting richer?

Despite increases in net wealth even for the most poor in a society, it looks like increasing inequality can destabilize the economy source and potentially give rise to warfare and revolution source.

Rising income inequality makes socio-economic mobility more challenging source and is linked to negative social phenomena such as depression, conspiracies, anti-immigrant sentiment, and rise of intolerant populism.

Potential Solutions to AI-induced income inequality

The polarisation of the job market seems like a real issue that we're going to face in the future, but it does seem like a tractable problem. There are some things we could and should attempt;

  • Anticipate mass displacement of workers and put in place general training programs that teach people how to use new technology to their advantage.
  • Revive highly-skilled artisanal trades like stone masonry and woodwork, in the hope that a rich, highly-automated society will shift towards demanding high quality, humam-made things.
  • Improving welfare programs to support those most affected.
  • Overhaul the education systems to better prepare future generations for an AI-dominated era.
  • Use our resources now, before it's too late, to figure out what fair distribution of AI's benefits means source.