Artificial intelligence (AI) has crossed a threshold. “In the past five years, AI has made the leap from something that mostly happens in research labs or other highly controlled settings to something that’s out in society affecting people’s lives,” says Michael Littman, chair of the One Hundred Year Study on Artificial Intelligence, hosted at Stanford.

It’s easy to see what he’s talking about: The technology’s impact can be seen introducing automation, driving efficiency gains and enhancing productivity, creating new jobs, and reducing risks associated with cyber-threats and fraud. During the pandemic, AI enabled more effective testing for Covid-19 and faster vaccine development, and helped manage grocery supply chains and tailor lessons for individual students affected by remote schooling.

As AI expands into more and more facets of our lives, there is also more scrutiny on who’s developing it. To ensure that AI is built and deployed in ways that are ethical, work for broad and diverse populations, and protect civil rights and freedoms demands that its development is not captive to the ethical judgments, design choices, and application priorities of a narrow population of developers. It’s understandable that AI developers are influenced by their own world views, which, in turn, guide them in their selection of  applications, datasets, and training of algorithms. These world views are shaped by factors such as gender, race, ethnicity and geography; therefore, it is even more critical in AI than in other tech fields that the talent pool is diverse and inclusive as it has such a profound impact on the product itself.

We know that there are big gaps to close when it comes to improving the diversity of the talent pool. Research done by our Digital Planet team at Tufts University’s Fletcher School shows that 17% of the AI talent pool is female as compared to 27% in STEM overall. Black workers constitute only 11.8% of the AI tech workforce, according to the Center for Security and Emerging Technology at Georgetown University.

A major barrier to addressing this problem is that skilled workers and other key resources cluster around a small number of urban hubs. This causes a “de-democratization” of AI research and development, limiting it to a handful of elite companies and universities and, therefore, geographies. For example, as a Brookings study points out, the San Francisco Bay Area is so highly concentrated in AI-related activity that about a quarter of all U.S.-based AI conference papers, patents, and companies are there — and it has four times the AI activity of other top cities with AI clusters.

For all the reasons we highlight above, companies that develop and apply AI need to recruit from a wider set of sources. This might require opening offices in multiple locations to ensuring that recruiting teams know where to go to find diverse talent. The good news is that, while not as large a cluster as the San Francisco area there are other AI-rich urban clusters for companies to consider.

In this article, we highlight the 50 cities with the largest AI talent pools worldwide and evaluate them using a framework that we have developed at Digital Planet: TIDE (for Talent pool; Investments; Diversity of talent; Evolution of the country’s digital foundations). These factors collectively give companies a way to prioritize their AI talent sourcing choices by scoring the different locations on the concentration, quality and diversity of the AI talent pool. Our data on talent derives from the SeekOut database and location analyses of AI professionals active on social media, while investment data derives from venture capital data by city. The diversity measure combines several factors: the proportion of female AI workers, racial diversity and migrant acceptance along with the cost of living in a city. Digital evolution draws from the work we reported on earlier in the HBR.

The top 50 AI cities as measured by the TIDE framework are:

There are several implications of this analysis.

While San Francisco leads in AI talent worldwide, there is a wider pool from which to draw. We find that in terms of AI talent concentration alone, the Bay Area is bigger than the combined total of the next two U.S. cities, Seattle and New York. But, its position is also probably overstated. There is a global talent pool that extends well beyond the U.S., including in the developing world. For instance Bangalore, India — the fifth-ranked city for diversity among AI workers, as measure by our TIDE framework — has the world’s second-largest AI talent pool.

The AI landscape is dynamic, especially in the developing world, and these changes can significantly transform the hotspots list over time. Brazil and India, for example, have cities on this list of 50 and are hiring three times as many AI workers as they were in 2017. This is a rate of growth that matches or exceeds that in the U.S. Almost 30% of scientific research papers from India include female authors, double the proportion of female authors in the U.S. and UK. Meanwhile, the Chinese Academy of Sciences is the top publisher of AI research, with Tsinghua University and Peking University close behind. Companies should consider expanding their hiring programs in these countries, wherever possible, to enhance the diversity of their AI developer pool.

Remote working can help spread AI activity more widely. With the growing acceptance of remote work, there are greater opportunities for cities that have a lower cost of living to become more important players in AI. Cities on our list of AI hotspots that are in the developing world, such as Hyderabad, Bangalore, Jakarta, Lagos, Nairobi, Mexico City, Buenos Aires, and Saõ Paolo, score favorably on cost of living, which could be a powerful draw for diverse talent, especially those from disadvantaged backgrounds. A key consideration is that the potential of these cities is constrained by the state of their digital evolution and infrastructure.

Cities in Europe such as Tallinn, Madrid, Barcelona, and Berlin are in the sweet spot of relatively lower cost of living and a high state of digitalization. To realize their full potential, all of these cities need to improve in their acceptance of migrants. In the meantime, companies can identify the cities from these regions that can work well with their existing corporate presence, customer needs, organizational footprint, time zones and team structures — and determine recruiting and retention strategies accordingly.

Companies aiming to narrow the gender gap should consider cities with higher proportions of female AI talent. With slightly more than 25% women in its AI talent pool, Tel Aviv comes out on top globally on gender diversity. Along with the North American cities, and Edinburgh and Buenos Aires, it joins 13 other cities with at least 20% women in its AI talent pool. Boston performs the best in the U.S., with the San Francisco Bay Area at number two. Companies across the board have made commitments to improving the gender balance of their tech workforce. Many ideas have been proposed on how to address the current imbalance and its negative impact, especially in AI; devoting recruiting resources to cities with greater female representation in the talent pool can be a crucial part of the solution.

Company recruiters considering U.S. cities, must balance a tradeoff between size of the AI talent pool and its racial diversity. Atlanta, Washington D.C., and Los Angeles — while ranked lower, at numbers 6, 10, and 11, on our overall rankings — have higher proportions of Black and Hispanic AI talent in the U.S. as compared to the top-ranking U.S. cities overall.

Companies can encourage national governments to consider immigration policy reforms that help with building more diverse AI talent pools, especially in the university systems. Cities in the EU present a critical case in point. In general, the EU lags the U.S., UK, and China in the number of STEM undergraduates relative to the population; meanwhile, a large proportion of its STEM graduate degree holders and PhDs from the EU seek jobs elsewhere. The EU labor market overall attracts a relatively small share of international AI talent. The EU has an additional set of challenges when it comes to faculty: The prime academic institutions for AI talent are suffering from faculty shortages. In Germany, for example, one STEM professor advises 90-100 students on average. Greater acceptance of migrants can help build up the EU’s AI capacity and outputs.

Companies can collaborate with local city governments to nurture vibrant hubs attractive to a tech workforce. An important facet of a robust talent strategy is to go beyond recruiting by investing in retaining the best recruits. Investments by companies and governments to build supportive civic institutions, urban amenities, incentives and a city’s brand can go a long way in building loyalty. Such public-private collaborations could also compliment a mayoral and city government’s public campaigns to attract investments, high-earner workers and jobs. While such investments are particularly attractive for companies that set up offices in the cities, with remote working, it is also feasible that cities become magnets for employees who work for a company based elsewhere.

For a general-purpose technology, such as AI, to succeed, it must embrace applications, data, and, most crucially, talent from multiple contexts and sources. Currently, there is a shortage of AI talent and companies need to think expansively about their where to recruit AI talent. O’Reilly’s 2021 AI Adoption in the Enterprise reports that a shortage of skilled AI workers and difficulty in hiring tops the list of AI challenges, while the three most in-demand skills on are AI related.

There is no question that companies must extend their searches beyond the “usual” cities — the classic tech hubs — from which to recruit. There is good news: The search for AI talent is also a tale of 50 cities, and possibly even more. The hotspots noted here benefit from promoting their cities as the AI industry widens. But all of us as consumers of AI products and enabled services also benefit from a technology that draws on a breadth of sources to produce new tech applications that are free of bias, protect freedoms and work for everyone, everywhere.

The authors are grateful to Paul Trueman at Mastercard and Joy Zhang at The Fletcher School at Tufts University.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *