Future and emerging technologies: Economic impact of artificial intelligence

Event report

[Read more session reports from the WSIS Forum 2018]

Mr Catalin Marinescu, head, corporate strategy division, ITU, started the session by noting that artificial intelligence is moving into everything from finance to health. He highlighted that if yesterday the discussions were about the need to connect people, today the debate is about everything becoming intelligent. Marinescu screened a video showing AI contributions to the realisation of the United Nations sustainable development goals. He then asked the first panellist, Mr Jacques Bughin, director and senior partner, McKinsey Global Institute, to talk about the possible impacts of artificial intelligence in different sectors of society.

Bughin noted that today when we speak about artificial intelligence there is always one group stressing possible negative impacts in areas such as security and employment, and another group praising all the potential and benefits that we will be able to enjoy. He said that he would like to frame his presentation another way, by presenting what we really know about AI today, what we can estimate from what we know today, and using these elements propose some “calibrations” about the future.
Burgin said that actually we know that:

1. AI is a reality that is only at the beginning, but evolving fast and steadily, including in core areas such as machine learning.
2. Economic aspects are influencing AI development, particularly because technology is getting cheaper and its deployment impacts directly in productivity growth.
3. There is no AI without digital, and this connection highlights the importance of addressing the digital divide issue.
4. Automation and the risk of cognitive skills-based activities being replaced by robots will cause some jobs / occupations to be ‘killed’.

From this Bughin inferred that the economic aspects will continue to support AI development, causing a competitive race to push AI into the automation of cognitive skills-based activities, and the service sector. Based on simulation scenarios, he noted that AI will impact productivity growth at rates not observed before, when compared to changes caused by other technologies. He stressed the importance of design policies to shape the way this growth will be distributed in order to avoid increasing inequality. Bughin finally stressed the need to understand that AI is here to stay, and that all actors must be ready to create policies which develop the skills needed in the future.

Marinescu thanked Bughin for his presentation and invited Ms Irmgard Nübler, senior economist, the International Labour Organization, to express her views about the impact of AI in the dynamics of jobs destruction. Nübler started by noting that technology deployment has always impacted jobs in the past, and whenever a new technology is developed this happens. She then pointed that the quest for increasing productivity and the use of innovation and technology supporting this process always leads to jobs being ‘killed’.
She stressed that in order to address these destructive waves it is necessary to promote economic and social restructuring, particularly promoting social capabilities able to prompt the creation of new jobs and markets. Nübler also pointed out that in order to address the challenges caused by AI development, a collective effort must be made to promote the creation of the knowledge base and the social capacities necessary to embrace the potential of the AI revolution.

The moderator than explained that due to technical issues the panel was unable to connect with Mr Neil Sahota, IBM master inventor, IBM Watson Group, and asked Prof. Mary-Anne Williams, director, disruptive innovation, University of Technology, Sydney, to talk about the risks involving AI development.

Williams started by noting that AI is impacting all social and economic aspects of our lives, including the way we interact with others. She said that there is nothing that AI cannot do, and screened a video illustrating this. She then pointed to some of the risks of AI being weaponised or used to violate established rights, such as privacy. She said that these risks, according to the United Nations, should be carefully analysed from the perspective of human rights, peace and security, and sustainable development. She concluded by underlining the need to develop policies and regulatory frameworks to ensure safeguards against these risks.

By Joao Araujo Monteiro