A recent UN Agency publication has attempted to quantify the impact that Artificial intelligence will have on Global labour markets, and project around 88 million jobs will be augmented or automated by AI Agents.
The paper does not anticipate sudden and massive job destruction from AI systems like GPT, rather a restructuring of the division of labor. In effect generative AI, in the form of digital AI agents, will reduce the number of cognitively simple and repetitive jobs allowing people to add value to more nuanced challenges. Polanyi's Paradox will prohibit full AI Agency and automation for a time still.
Polanyi’s Paradox is the term first coined by the Hungarian-British polymath Michael Polanyi when he discovered that: “We know more than we can tell”, i.e., the knowledge of how the world functions is, to a large extent, beyond our explicit understanding.
The paper analyzes the potential exposure of different occupations and tasks to automation or augmentation by generative AI systems (AI Agents) like GPT. It finds that overall, augmentation effects are likely to be more common than full automation of jobs, though automation potential is higher in clerical and some para-professional roles.
The pace of deployment and adoption of generative AI systems is highly uncertain and will vary across countries. Constraints like infrastructure gaps, costs, regulations etc will affect timeframes and new compensating job growth is not accounted for and may offset declines in exposed occupations.
The impacts are not instantaneous but rather gradual changes in the composition of employment over decades with historical precedents show new technologies can take many years to diffuse. The paper does not anticipate sudden and massive job destruction from AI systems like GPT, rather a restructuring of the division of labor.
We cannot know now what kinds of opportunities this will lead to, and how this could open up jobs to more interesting and engaging work. This could be similar to the impact of new technologies like the first printing press which created new skilled trades and consumer markets, via a Renaissance, in ways that were impossible to predict initially. For high-income countries, 5.5% of total employment (~24.6 million jobs) were found to be in occupations with high automation potential from AI Agents. In upper-middle income countries, this figure was 2.5% (32 million jobs), for lower-middle income countries it was 1.2% (~30.5 million jobs), and for low income countries it was just 0.4% (~1.3 million jobs) of total employment.
The job functions most likely to be augmented or automated by AI Agents are:
Clerical support roles: clerical workers had the highest share (24%) of highly exposed tasks that could potentially be automated. This includes jobs like secretaries, typing clerks, data entry operators, and other administrative support roles involving routine information processing.
Certain customer service jobs: Customer service roles involving repetitive communications like call center workers, reservation clerks, client information workers etc. were also found to have high automation potential.
Accounting and bookkeeping clerks: Financial clerical roles doing routine data entry, record keeping and calculations were identified as prone to automation.
Bank tellers and cashiers: Client-facing finance roles at banks and retail stores could decline with automated self-service options.
Production line machine operation: Although not highlighted in this paper, routine machine operating jobs remain susceptible to automation by AI and robots based on prior research.
Data collection jobs: Surveys, polls, census-taking and similar data gathering roles could be impacted by generative AI's text and speech capabilities.
The common theme is that procedural, rules-based and repetitive information processing tasks are most exposed to automation. However, the paper stresses that augmentation, where technology automates some but not all tasks, is a more likely outcome than full automation of occupations. Jobs requiring advanced social skills, creativity, or dexterity were found to have lower exposure.
We are at an incredible time to be alive, with new technologies that we are only just beginning to harness in practical ways that will make our lives significantly better. Dropping the price of cognitive labour to almost zero will change society fundamentally, and in the best possible ways.
About the Author
Rudy Nausch is a tech leader and consultant specializing in AI technologies and AI Agents.
I welcome discussions on how AI can help shape your organization's future and how we can develop AI Agents together.
Visit daimonic.ai for more.
Commentaires