Skills for AI Adoption

Background

As technology use becomes increasingly pervasive, Canada's economic performance could be greatly enhanced by the rapid adoption of new technologies - AI and machine learning in particular - into existing businesses. However, as it stands Canada is a laggard when it comes to the adoption and implementation of technology.

As a general purpose technology, AI has the potential to improve Canadian productivity across sectors and is estimated to deliver over CAD$17 trillion of economic impact worldwide by 2030. However, like other general purpose technologies, such as electrification, these gains will only be realized if Canadian firms have the necessary complements in place, including structured data, redesigned workflows, infrastructure, leadership, and talent.

A strong supply of highly skilled workers to design, implement and work with technology is a key enabler of technology adoption. This is perhaps best exemplified by the unprecedented investments in human capital made by the U.S. in the early 1900s, which drove American economic prowess during this period of rapid technological change.

Unfortunately, talent deficits are already likely delaying the adoption of this critical of technology into thousands of Canadian companies. According to a recent survey, despite 89 percent of Canadian executives agreeing that AI will create substantial near-term value, only 34 percent have devised an AI strategy. Another study revealed that only around 20 percent of over 3,000 AI-aware C-level executives across the world have adopted AI at scale in a core part of their business.

Like other technological developments, AI will require individuals with technological capacities, as well as uniquely human skills that augment technology. As modern AI becomes more effective at making predictions, it will increase the demand for, and value of, human judgment across the economy. Investments in educating machine learning experts is increasing, but the complexity of AI systems also requires workers with complementary skill sets. This includes individuals who will bridge the gap between AI technologists and business professionals, and those who will ensure that AI systems are functioning appropriately and that any unintended consequences are addressed. The adoption of AI and the incorporation of relevant talent will also require new management skills and processes to effectively harness the value of this technology. The acquisition of these skills and integration of effective management processes are essential to ensure that Canadian firms are able to adopt AI at scale.

Project Description

With the Brookfield Institute for Innovation + Entrepreneurship, I am proposing a study to closely examine how Canadian companies can position themselves to take advantage of artificial intelligence (AI) for enhanced competitiveness, with a focus on the kinds of talent and business processes AI systems require to be adopted at scale.

Through this research project, we are seeking to identify the barriers to AI adoption associated with the kinds of talent, business processes, and organizational structures companies need to adopt AI systems in Canada. It is our goal to help policymakers, educators, and business leaders more effectively prepare for AI adoption so we as a country can more effectively capture the value of the technology.

We will seek to answer the following research questions:

  1. What are the skills and competencies that Canadian businesses require to successfully adopt AI tech and systems?

  2. What organizational structures and business processes are required to integrate this talent and successfully adopt AI?

  3. What skills do organizations struggle to find?

  4. Are organizations aware of the kinds of talent they need to adopt AI systems effectively?

Outcomes of this research will include:

  • A basic, preliminary typology of common AI skills and competencies and a typology of AI job types, (e.g. including technical categories like AI/ML developers, data analytics / statistics, etc; but also more generalist expertise like AI business strategy, project/product management in AI, change management, evaluation and management of considerations related to legal, PR/marketing, ethical/public trust implications of AI, etc etc).

Next Steps

We are ready to go! We are just seeking funding partners to get this awesome project off the ground!

Geography of Skills

Background

Anyone that sits within a three desk radius from me knows that I have a professional crush on David Autor. But I was recently struck, not by one of his papers, but rather a presentation he made as the keynote for the American Economic Association Annual Meeting  — which I assume gets as crazy as we all think it does.

In this presentation, Autor shows that the trend of job polarization over the past three decades has hit workers in cities without a college education particularly hard. Cities in the US once had a plethora of high and middle-skilled jobs, all of which enjoyed increasing urban wage premiums, meaning they earned more money in the city then they would in rural areas.  However, as middle-skill jobs began disappearing from cities (and everywhere really), workers without a college education reoriented towards low-skilled service jobs  —  contributing to a decline in these urban wage premiums. Meanwhile, college-educated workers in cities filled the growing demand for high-paying professional and managerial roles.

The labour market is polarizing

Between 1970 and 2016 in the US, job growth has been largely concentrated in high-skilled occupations, such as professional, technical and managerial work. At the same time, there has been falling employment in middle-skilled, more routine-oriented occupations (such as manufacturing jobs, clerical positions, and sales). Low-skilled occupations, such as personal service work in health care, hospitality etc… experienced little to no change in aggregate.

As these changes unfurled, college-educated workers moved into high-skilled jobs, while non-college educated workers relocated from middle-skilled production and office jobs into low-skilled service-oriented jobs.

For more information on why this may be occurring see: Autor, Levy, and Murnane (2003), Goos, Manning, and Solomons (2014), and like 30 others! Also for job polarization in the Canadian context see: Green and Sand (2015).

Cities rock for college-educated workers, they rock WAY less for everyone else

Autor shows that cities are increasingly becoming the haven for high-skilled workers, while becoming more and more bleak for those without a college education.

In the 1980s, cities used to be home to a large number of both middle and high skilled work, providing opportunities for college and non-college educated workers alike. This trend changed. Between 1980 and 2015 there was a sharp decline in middle-skilled jobs, requiring less than a college education in urban areas. Meanwhile, high-skilled jobs, requiring a college education, continued to grow in urban areas. Low-skilled work is currently as prevalent in rural as it is in urban areas.

As these relatively lucrative middle-skilled jobs in the city disappeared, non-college educated workers likely relocated to service jobs, which pay less. As a result, between 1970 and 2015, the urban wage premium for non-college educated workers in the city declined drastically, especially after the 2000s. This is most pronounced among young workers 25 to 39. Meanwhile, the wage premiums for college-educated workers in cities increased after the 2000s.

Project

This one is much less fleshed out, but I am looking to examine how Canadian cities and rural areas have evolved in terms of skills composition, and what this has meant for wages for individuals with lower levels of education. Right now I am looking to link national occupation codes from the 2001, 2006, and 2016 censuses so I can start to do some research.

If anyone is working on anything similar or wants to help out, please reach out!