Solutions Paper - Solving Employment in the Light of Automation

by Geneva Macro Labs May 28 2017 updated Jun 19 2017

GeM Labs solution space to tackle labour market challenges that arise from the rise of automation and artificial intelligence.

This document is thought to stimulate further discussion and is not a final presentation of the outcomes.

A publishable version will be created after the end of July to allow ample time for everyone to contribute their thoughts.


Automation and robotisation have created significant insecurity and uncertainty regarding their societal impact. Will we still have enough jobs? Will inequality increase? Will artificial intelligence steer and dominate our lives?

So far, automation does not seem to have generated massive adverse effects on labour markets. At the same time, it has not generated many benefits in the form of faster productivity growth, either (Gordon, 2016).

Job destruction has declined but inequality has increased: Job destruction has declined but inequality has increased

What has increased, however, is inequality, especially within countries. Job polarisation has led to a fall in middle-class jobs and increased both jobs for high- and low-skilled people (Autor, 2016, Nubler, 2017).

Solution Spaces

Inequality and skills

New technologies open opportunities for entrepreneurs and workers to venture into new areas and jobs. At the same time, it creates the risk of leaving behind those who are unable or unwilling to adapt to these new openings. In previous decades of technological change, inequality has increased - if at all - only temporarily and new chances translated into prosperity shared by all. The question: Will this time be different?

Emphasising collective intelligence

Skill is hard to define. Difficult to reduce skill to something individual - can reside within a group. Co-leadership also means co-responsibility and thus it is hard to reduce things to individual levels.

Redefining skills and jobs

There are personal, cultural & collective skills. Redefining skills and work will drive different policies which are essential for structural change.

Ensuring moral alignment of autonomous tech

Need for AI designers and mechanical engineers to develop moral standards. AI should be a collaboration project, not a competition.

Job loss and supporting transitions

Managing change

How can an ageing society cope with so much change? Is a younger society actually more apt at dealing with ever-faster developments?

Evolving education systems

Education is important but not necessarily traditional 12-year education, rather focus on core skills necessary to move ahead continuously. Information is free, coding is accessible to anyone.

Rethinking welfare

Rethinking safety nets

As work evolves at higher rates of change between sectors, locations, activities, and skill requirements, many workers will need assistance adjusting. Many best-practice approaches to transition safety nets are available and should be adopted and adapted, and new approaches considered and tested. If automation (full or partial) does result in a significant reduction in employment or greater pressure on wages, some ideas such as universal basic income, conditional transfers, and adapted social safety nets could be considered and tested.

Development potential for emerging economies

Fragile or failed states are very different to Middle-Income Countries, therefore, we need more distinction here.

Leapfrogging to level the playing field

Focussing on new technologies with initial costs and big marginal gains will allow narrowing gaps if done right.

Global redistribution

Is essential but hard to achieve. This change might help spur a new type of governance.

Focussing on job creation

Capturing the productivity benefits of technology. These can be harnessed to create the economic growth, surpluses, and demand for work that create room for creative solutions and ultimately benefit all. Accelerating the creation of jobs in general through stimulating investment in businesses, and accelerating the creation of digital jobs in particular - and digitally enabled opportunities to earn income. This includes new forms of entrepreneurship to move toward more self-employment. There's a lot of development potential in the service sector. Generally moving towards self-employment across all sectors.

Smart governance and international coordination

Tractably optimising international coordination

It is hard to overstate the importance of this option. Even though it will be difficult to achieve, there is no way around global governance with issues of global scale. The sooner we experiment with possible approaches, the sooner we will find out how to tractably improve global coordination. Coming up with actionable change is one thing - being listened to and having it implemented is a whole different question.

Rethinking nations

The coming wave of international challenges might cause enough of a stir to become an opportunity to reshape mindsets and redefine what a nation is and what international bodies do and what kind of responsibilities they carry respectively.

Creating new institutions

A new series of bodies specialised in different areas of the developments will be necessary to complement existing organisations. These institutions should focus on global, broadscale developments, such as education systems; R&D; healthcare; safety nets, tax systems.

Researching unknowns

It is not clear what should be done nor how it can be coordinated. We need to do R&D in many fields: (i) jobs (equivalents) in an automated society; (ii) which skills will stay in demand; (iii) regional idiosyncrasies and how to account for them; (iv) prioritisation research to understand what to tackle first; (v) effective approaches: institutional, economic or

Focussing on consequences

Researching the consequences of automation and figuring out how to optimise the job creation process is important. First ideas voiced were: (i) aligning economic incentives with public interest; (ii) changing our objective/identity as a society; (iii) changing the definition of a job

Creating proactive policy

We need policies to ensure that one individual doesn’t gain too much power through automation.

Exploring public-private partnerships

Investments in digital infrastructure would unlock digital benefits in many economies, both developing and developed; public-private partnerships could help address market failures. Furthermore, through tax benefits and other incentives, policy makers can encourage companies to invest in human capital, including job creation, learning and capability building, and wage growth. Driving training also could be key here. Companies face gaps in skills they need in a more technology-enabled workplace. They could benefit from playing a more active role in education and training, including providing better information about needs to learners and the education and training ecosystem and creating better learning opportunities.

Questions raised at the event




Development potential

Actionable Steps

1. Tackle inequality, joblessness will take care of itself

  1. Taxing robots
  2. The impact of technology on the welfare state?
  3. International tax cooperation
  4. Breaking up monopolies created by AI and technological leaders

2. Skills for the world

  1. Anticipating skill needs
  2. Adapting education
  3. Focus on those who lose their jobs, prepare the transition

3. Make robots your allies, not your enemies

  1. Leapfrogging possibilities for developing economies
  2. New opportunities in services industries
  3. Need to create a level playing field across industries and countries


Konrad Seifert

Development potential for emerging economies

I believe that here one key factor would be gathering a lot more data on the voluntary sector as Machine Learning and AI thrive on data, however, most data is generated by things that necessitate substantial investments. If we do not close that gap intentionally, it will only widen, as more wealth means more tech which means more data which means more wealth etc.

Maybe a good first step to closing global inequality and really capitalising on the potential of automation would, therefore, be the improvement of systematic collection of trustworthy data on a large scale. I am unsure how we could achieve this though.

Konrad Seifert

Ensuring moral alignment of autonomous tech

There are a few research centers looking into this kind of thing but it seems potentially pointless if we do not achieve global cooperation on these things. Thus, I believe, a priority should be policy strategy and prioritisation research for the international organisations.

Konrad Seifert

It is not clear what should be done nor how it can be coordinated\. We need to do R&D in many fields: \(i\) jobs \(equivalents\) in an automated society; \(ii\) which skills will stay in demand; \(iii\) regional idiosyncrasies and how to account for them; \(iv\) prioritisation research to understand what to tackle first; \(v\) effective approaches: institutional, economic or

A first step might already be to open up data silos of which there are plenty within IOs alone already.

Konrad Seifert

Global redistribution

Redirecting aid/philanthropical support towards data gathering programs in hope Machine Learning will be able to help solve problems better?


Automation favors those who own the capital. If I squint, I can see skills and ability to work (labor) kind of looks like capital that you can't easily trade for new ones. Automation is capital that has some of the same properties of skills and ability to work but is more liquid than human labor. All these solutions seem like different forms of redistribution. Do we redistribute the skills necessary to build and run automated systems. Do we redistribute the data necessary to train these automated systems. Do we redistribute ownership of already established automated systems?