Redistribution is key to the future of the workforce

Outlook

Twentieth century data supported the view that machines could only replace routine tasks. A 2003 paper by Autor, Levy and Murnane became the ‘ALM’ consensus, assuming non-routine tasks could not be automated as they relied on knowledge humans could not explain to robots.

Changing our understanding, machine learning capability has since progressed faster than human capability. In 2016, AlphaGo, Google DeepMind’s artificial intelligence computer programme, beat the human Go champion with a new move. Algorithms designed the widely praised interior of the newly opened Hamburg Concert Hall.

Consumers benefit from greater accessibility and affordability. Potential for ageing societies is vast; a 130,000-image algorithm is more reliable diagnosing skin cancer than 22 doctors. However, redistribution is critical to minimising disruption.

Impacts

• One certainty of the race between workers and machine owners is inequality, making structural reforms and redistribution policies crucial.
• London School of Economics Professor Leslie Willcocks estimates that for every 20 jobs robotics destroys from 2017-27, it will create 13.
• Legal and regulatory regimes will struggle to keep up, and education shortfalls in advanced countries will intensify competition for talent.
• Premature deindustrialisation is worrying for developing countries with large, young workforces that are industrialising from agriculture.