The mention of AI and machine learning in earnings reports has risen seven-fold from 2015 to 2017, but conversations in business circles tend to focus on what this means for costs and margins. The WEF's latest estimate is that machines and algorithms will perform 42% of total task hours by 2022, up from 29% in 2017, and increasing to 52% by 2025. Benefits to efficiency are indisputable, and many applications will have positive effects on emissions, energy consumption and use of material inputs. But the human costs are less clear.
Most jobs retain functions than cannot be easily automated, and automation sometimes improves safety and working conditions, but the reduction of demand for many types of labour and the need for different skills in future jobs will adversely affect the livelihoods of people across the world. Many western countries are still dealing with the socio-cultural consequences of large-scale deindustrialisation in the 1980s and 90s. Governments and business need to start thinking about automation now to avert similar results on a global scale. The challenge: How might employers prepare for the human cost of automation?