Interpretable AI in HR tech: the only way of understanding machine decisions

Largely spurred by the success of ChatGPT, governments worldwide are finally attempting to find a way to regulate AI.

Welcoming Trond Henning Olesen as our new VP of Customer Integration and Solution Sales

Trond Henning Olesen

We are excited to announce that Trond Henning Olesen will be joining as our new VP of Customer Integration and Solution Sales, based in San Francisco. He will be responsible for all accounts in the Americas, EMEA and Asia.

Trond is a highly experienced strategist, technologist, and startup enthusiast. Leveraging over 20 years of global experience in leadership and sales in the tech industry, as well as a PhD in Computer Science, Trond brings an impressive track record of successfully building customer-facing teams, launching new ventures, and delivering operational impact.

Throughout his career, Trond has built businesses from startup to successful IPO, achieved top growth, turnarounds and high customer satisfaction in diverse market conditions. He has also delivered major accounts and managed complex large projects across the globe, as well as effectively leading teams to bring about fundamental changes and improvements in strategy, process, and customer focus. With his extensive technical and business expertise, Trond has consulted for companies such as LinkedIn and Purisma, personally coaching C-level personnel and assisting them in improving their organization, processes and people. Most recently, he co-founded and served as CTO of Silicon Valley startup VeraScore, managing the technical team, participating in development and being the technical lead on all sales efforts.

Trond is enthusiastic about the highly performant AI-driven job matching technology and labor market solutions offered by Swiss-based to businesses and government institutions around the world. In these times of major structural changes of the labor market, Trond is energized by the opportunity to work with global clients to provide them with perfectly tailored digital solutions for effective talent and labor market management.

“With his deep blend of a strong technical background and expertise in strategy and customer success, Trond is an outstanding addition to our team,” states Stefan Winzenried, CEO of “As we continue to deliver quality, cutting-edge solutions, Trond will accelerate JANZZ’ growth and strengthen our mission to better serve our clients. We are thrilled to have him on board.”

Wanted: healthcare workers – but why aren’t these jobs being filled?

Smiling female nurse holding senior woman’s hand. 

Despite improvement, there will still be a significant gap between supply and demand of healthcare staff by 2029 in Switzerland, according to the national 2021 report on future healthcare staff needs, published by the Swiss Health Observatory in September.

The report estimates that by 2029, the personnel demand in the healthcare sector may rise to 222,100. Compared to a base number of 185,600 recorded staff in 2019, an additional number of 36,500 staff will be required. To fill these additional positions, as well as to compensate for those who retire and leave the industry early, presents an enormous challenge to the next generation of health workers. The report further stated these jobs remain difficult to fill mainly due to demographic and epidemiological developments.

The growing need for qualified healthcare professionals is not specific to Switzerland, a trend can be seen around the globe, and it is not a recent trend. Back in 2016, World Bank published the research on Global Health Workforce Labor Market Projections for 2030. It predicts that global health workers’ demand will increase to 80 million, with a supply of 65 million health workers over the same period, resulting in a global shortage of 15 million health workers.

In a previous article, we have also discussed the shortage of healthcare professionals in the Swiss labor market (see Switzerland 2030: The risks and opportunities of digitization). Today, due to the effects of the global pandemic, it might have impacted the situation negatively. The pandemic has brought renewed attention to the frightening shortage of health workers, but remember, this extraordinary situation existed long before the Covid-19.

The same pattern is clearly evident in some other industries and occupations (see Free movement of skilled workers in the EU and beyond are more important than ever), with fewer and fewer young people willing to learn such skills. It is a basic problem and therefore, needs some fundamental changes to effect sustainable action. Share with us your thoughts and let us know your opinion on this topic.

Follow-Up on Equal Pay, or; The monster in our closet that we all ignore

This is a follow-up post to our last article on the Gender Pay Gap (GPG), in which we suggested that focusing on a pay gap based on gender is not enough and that shifting the focus to the concept of performance would be useful. As a kind of continuation, we turn here to the topic of fast fashion and discuss this ubiquitous ‘monster’ in all our closets, including from the perspective of equal pay.

Lack of initiative despite well-known problems
We all know it,  » Read more about: Follow-Up on Equal Pay, or; The monster in our closet that we all ignore  »

Equal Pay – Let’s not fall into the gender pay trap…

For some, Equal Pay Day marks the annual ‘free labor of women’, for others it is an ‘ideologically motivated lie’ – the truth is even more complicated. This article highlights the uselessness of many of the statistics surrounding ‘equal pay’ and explains why a focus on gender falls short when it comes to pay equity. Instead, we would do well to place more emphasis on actual, individual performance.

Is a focus on gender even justified?

First of all, let’s take a look at the terminology. Equal Pay Day (EPD) is generally understood as an international day of action for equal pay between women and men. The day is intended to draw attention to the so-called ‘gender pay gap’ (GPG), i.e. the gender-specific wage gap between the average gross hourly wage of women and men. The EPD is celebrated individually in each country, symbolically on the day up to which women would work for free if they were paid the same amount as men across society as a whole. In Switzerland, the EPD falls on February 20. So much for the definition.

Let us now turn to the methodology used to measure and prove the GPG. Two statistical approaches dominate here. Either the gross wages of women and men are compared by means of a regression analysis, without taking into account objective factors such as level of education, qualification, age, professional experience, overtime hours or the exercise of a management function. In the case of Switzerland, this led to the following result in 2018: Women earn on average a striking 19% less per hour than men [1].

Due to the fact that objective factors were neglected in the analysis this result is, however, questionable. Such a calculation literally compares apples with oranges. The other approach used concerns the calculation of the so-called ‘adjusted wage differential’, in which a distinction is made between ‘explained’ and ‘unexplained’ differences in pay. ‘Explained’ are those pay inequalities that exist due to objective reasons such as full-time/part-time employment or work experience. The ‘unexplained’ share refers to the pay gap that is potentially attributable to subjective inequality based on gender and amounted to 8.1% in Switzerland in 2018 – still a considerable, but nevertheless notably smaller figure than the result of the first calculation [1].

Although the first approach is now considered outdated, its results are regularly cited in the media to create politicized headlines (for example here) or to serve the often shouted slogan “equal pay for equal work.” But even the adjusted GPG omits important explanatory factors. For instance, comparisons are not made task-specifically but rather on an industry-wide basis, or projections are made nationwide despite very large regional differences in the cost of living. If all these relevant aspects were also included in the statistics, the explained difference would probably be even lower. A slightly different objection is that if such huge differences really existed despite equal qualifications, every profit-oriented company would actually have to hire only women, which definitely does not correspond to reality.

One real consequence of all this is a (media-driven) battle between the sexes, which unfortunately seems to lead nowhere except to diminished solidarity between men and women. A prime example of this is Switzerland’s Equality Act, which was amended in 2020. Since July of last year, the largest Swiss companies have had to monitor their wages for discrimination and fulfill their obligation to inform their employees about the findings. Apart from the fact that the law only obliges companies with more than 100 employees, which excludes a large part of Swiss SMEs, it does not provide for any sanctions [2]. Recently, the first evaluations of this federally mandated wage analysis were published – with surprising results for some. Only 5 percent of the analyzed companies do not meet the stipulated requirement, which on the one hand is below the granted tolerance threshold and on the other hand clearly contradicts the previous statistics of the federal government. Interestingly, the reactions of GPG critics to this result are either non-existent or skeptical [3]. One almost wants to say, because they do not fit their expectations or their world view…

The bureaucratic effort that is made today for the purpose of equal pay leads in the end not only to contradictory statistics and little real change. Moreover, the analyses also distort the representation of the actual economic situation of a non-negligible share of men who are below the average value of their gender, since male top earners statistically make all men appear to be better earners.

Focus should be on performance

The slogan “equal pay for equal work” quoted in the last section is therefore misleading, since the standard analytical methods used to examine gender pay inequality do not compare like with like. Another, even more important point in the discussion of equal pay concerns the concept of performance. Achievement can be divided into two components, the effort made and the result achieved within a given period of time. However, the first component is often equated with the second, or considered an equivalent substitute for the lack of the latter. An insistence on actually achieving results is sometimes even frowned upon as inappropriately exercised pressure to perform. And yet, in the end, it is the performance that matters. Let’s use an analogy to the world of sports: Recently, at the Summer Olympics in Tokyo, the on average smaller beach volleyball players from Japan played against the on average taller athletes from Germany. Although body size can definitely play a decisive role in this sport, neither the relative performance of the Japanese nor which team had trained x hours more, nor who had already participated five times more in the tournament counted for the final result, but only their performances in the match.

From the perspective of both fair and profit-oriented employers, the situation is similar. When it comes to salary, only the actual performance – measured by the contribution to a company’s income statement – of an employee counts or should count. Of course, this means that unchangeable attributes such as height, gender, ethnicity or social background should have absolutely no influence on a person’s salary. In radical economic terms, however, such a position would mean that even objective factors such as the type and duration of education would no longer have to be taken into account. Even if a completed degree means high education costs and temporary wage sacrifice, this qualification is sometimes not sufficient to outperform a naturally skilled or otherwise trained worker. If person A assembles twice as many parts in an hour on the assembly line as person B, then person A should also be paid more in the long run or be allowed better employment and advancement opportunities. Accordingly, fair hiring and compensation are ideally based on individually measurable and comparable performance – i.e., on the reason why a person is originally hired or should be hired. Thus, the well-known guiding principle can be rephrased as follows: Employers should not adhere to the slogan ‘equal pay for equal work’, but to ‘equal pay for equal performance‘. A deviation from this principle is therefore a genuine act of discrimination – including against women who perform measurably better and more consistently.

To return to the topic of equal pay in relation to gender, a marginal comment should be made here: If remuneration is to be based solely on a person’s contribution to economic success, the question arises at least as to why domestic work and child rearing in one’s own household/family, which is still mainly carried out by women, is not also considered to be work worthy of compensation. After all, it represents a significant pillar of our current economic and social system.  Or, if this work is ‘outsourced’ to a room attendant and a day care center, why are such important services are paid so low? We all tolerate this circumstance without a shred of bad conscience. This indicates that there would likely be no broad willingness to make a fair wage possible for people working in these sectors, for example by means of tax-financed subsidy payments. But more about this in the next section…

What counts as (equal) performance; or, The real problems

As alluded to, the principle of ‘equal pay for equal performance’ raises the question of what should actually count as equal – or rather equivalent – performance. Here, too, the comparison with the world of sports can be made: While the winner of the FedEx Cup golf tournament receives around 10 million US dollars in prize money, a ski jumper only gets one thousandth of this amount at the most for winning the competition. Both athletes perform exceptionally well (and entertain spectators), and yet the pay is set completely differently. Such gaps do indeed exist across the economy and cement specific sectors as high or low wage industries that often also correlate with gender. Of course, one can always appeal to women’s self-responsibility when it comes to career choice, à la ‘It’s your own fault if you choose a low-paying job that can be done part-time – no one is forced to start a family’. But we could also think about why so-called ‘female professions’ such as sales assistants, hairdressers or geriatric nurses are systematically paid less and whether this is fair. If the pandemic has shown us anything, it is that many of these professions are ‘essential’ for the functioning of our society. The prime case is nursing staff, who despite unpleasant working hours, stress and indispensable service provision earn much less than, for example, construction foremen. The real problem is therefore the wage inequality across sectors, which widens the gap between the rich and the poor and does not pay ‘equal’ performance equally.

For the individual, the consequences of choosing an occupation in the low-wage segment mean not only lower pay but also fewer opportunities for further training and promotion, as well as a lower pension in old age. But, to return to the GPG aspect: Discrimination can occur not only on the basis of gender, but along many lines, such as skin color, appearance, age, disability or religion. So why this focus on gender when solidarity and alliances can be formed much more broadly? On this point, one critic of the EPD suggests, “What would be really interesting would be a ‘social’ ‘Equal Pay Day’. […] The gender issue hides wealth and hides social inequality. The gender pay gap is […] a gender trap – a ‘gender pay trap’, so to speak” [4].

In summary, we as a society need to ask ourselves whether we are willing to accept the structural reasons for wage inequality as an unchangeable fact or whether we are willing to do our part to change the status quo. In terms of gender, one commentator summarizes that having children is not part of the plan in today’s economic model, as it often means being locked into a traditional role distribution and lower employment opportunities for women [5]. Here, a rethinking and a restructuring on different levels – individuals, large corporations, politicians – seems to be called for.

The examples of gastronomy and care can also be used to illustrate that social change requires everyone to take responsibility. Are we willing, provided we can afford it, to pay more for consumption in a restaurant or even higher health insurance contributions so that the waitress and the nurse are paid ‘more equally’ for their services, or do we prefer to be stingy so that we can keep more for ourselves? Do we elect policymakers who will advocate for better minimum wages in care, or do we think we’ve contributed enough by clapping from our home balcony? Or how about the next time you pick up your kid from daycare, you tip the caregiver a few hundreds to show that you appreciate their efforts and are aware of the wage disparities in the care sector?

At JANZZ, we are not satisfied with a couple of heart emojis posted on social media as extra pay for caregivers, but develop evidence-based solutions and have been using them successfully since 2010. Our job and skill matching solutions are fair and non-discriminatory and deliver completely unbiased results according to the OECD principles on AI. This ensures that the best performing candidates in all individual criteria receive the best match – regardless of employment status or other irrelevant characteristics such as origin, age, BMI or gender. This is one of the many reasons why we are a trusted partner for an ever-growing number of Public Employment Services (PES) around the world.

Want to take that first step to change the status quo and contribute to a more equitable labor market for all? Then contact us at or visit our product page for PES.



[1] Eidgenössisches Büro für die Gleichstellung von Mann und Frau. Plattform Lohngerechtigkeit. URL:

[2] Tagesanzeiger. Ab Juli müssen Unternehmen die Lohngleichheit kontrollieren. URL:

[3] Steck, Albert. Frauen werden kaum diskriminiert, wie die neue Überprüfung der Löhne zeigt. URL:

[4] Moser, Thomas. 2017. Ten Years Gender Pay Gap-Mistake – Ein Irrtum wird zehn Jahre alt. URL:

[5] Zufferey, Marcel. Die Mär von den unfairen Frauenlöhnen. URL:

Global Labor Market News: Despite strong vacancy recovery, long-term unemployment remains

A new analysis of online job postings from the UK confirms that there has been a strong recovery of job vacancy rates this summer, well above the same period two years ago. The data, coming from one of the largest online job search engines in the UK shows:

  • New vacancies are mainly in IT, construction, trades or warehousing and logistics with 330,000 in total, nearly one-third of all vacancies.
  • The numbers of vacancies in healthcare, nursing and social work are reaching 130,000.
  • The numbers of vacancies in both sales and hospitality are nearly 75,000.

Despite the increasing job advertisements, what raised more concerns, however, is that the unemployment rate remains high. The long-term unemployment rate keeps rising at the fastest rate in a decade and the unemployment claims are still double the size compared to the time before the pandemic.

Not only in the UK, the same trend can also be seen in many other rich economies. A report released by the OECD in July warned that rich countries may face sustained growth in long-term unemployment. Their data shows, by the end of 2020, the number of people in OECD countries who have been unemployed for more than six months is 60% higher than the pre-pandemic level. This is because the low-skilled workers who are most likely to lose their jobs in the early stages of the crisis do not have the skills needed to enter the most extensively recruited industries.

To deal with the situation, governments are suggested to develop more targeted employment retention programs to ensure that they do not support businesses that are unlikely to survive in the open market. Meanwhile, they should better cooperate with employers to support reskilling and job matching.

Not entirely a “she-cession” but globally women are the key to economic recovery

Advanced economies in Europe and North America are finally emerging after a year of COVID-19 lockdowns largely due to mass vaccinations, while populations in Africa and hard-hit South Asia and Latin America grapple with both vaccine access and labor informality. In a year mired by uncertainty, the economic and societal shocks of the pandemic impacted women and men differently—across the world, women were more likely to lose jobs, cut back paid hours worked, and became the default childcare providers in households.

In much of the world, women were left with no choice. In the United States, the pandemic erased the strides women had made in labor force participation rates since the 1960s.  By 2019, American women made up more of the workforce than men (approximately 50.04 percent of payrolls). [1]  Today, American women’s labor force participation stands at 57.4 percent which is lower than the pre-pandemic 59.2 percent in February 2020 and the lowest level since December 1988. [2] It is true that women’s employment suffered because many works in the services sector which witnessed hard hits in retail, healthcare and hospitality, while in developing countries women face high levels of informality which lack social safety nets to buttress the financial impact of a pandemic.

Last summer, an analysis by the McKinsey Global Institute showed that women comprise 39 percent of the global labor force but represented 54 percent of total job losses due to Covid-19. Many women’s choices disappeared as the pandemic created a childcare and education crisis with disruptions to everyday life.

Reactivating the economy with women in mind in a post-COVID world

Unemployment impacts men and women differently because society expects men to work and be the breadwinners, while women even with similar education levels as partners or husbands spend more time caring for the household. Perhaps it can be said that unemployment is shaped by gender, class, and social norms. While unemployment increased for those without university degrees or the ability to work from home, women and minority populations were disproportionately impacted. Even prior to the pandemic, gender inequality put women under greater financial pressure with unstable work contracts and less access to education and technology than their male peers.

The World Economic Forum’s Global Gender Gap 2021 report points out that the pandemic has increased the timeframe it would take to close the gender parity gap from 99.5 years to 135.6 years in terms of salary, education, and political empowerment. Overall, jobs provide people with earnings but more importantly they offer an identity that contributes to happiness and self-esteem. Covid-19 has left a void for women used to navigating a world of work instead of childcare.

The story of this pandemic is that women left the workforce independent of their education level or jobs held –at least in the United States, it appears this decision was based on whether children returned to in-person school. In Peru, children have yet to return to in-person schooling since March 2020, while other countries in the region have turned to hybrid models.  According to the International Labour Organization, approximately 13 million women in Latin America and the Caribbean witnessed their jobs vanish in 2020 due to the pandemic—with a regional total of 25 million women unemployed or out of the labor force. In developing countries, many women work informal jobs in vulnerable sectors usually earning daily wages with no sick leave or the safety net of unemployment insurance.

COVID-19 exacerbates digital biases and the gender pay gap

Additionally, women have dealt with digital biases (i.e., approximately 300 million fewer women have access to smartphones in low- and middle-income countries or 20 percent less than men according to the World Bank) in a world that overnight became more reliant on connectivity for survival coupled with the ever-present gender pay gap. Even today, in the United States women with school-aged children are slowly narrowing the gender divide and returning to work.

The pandemic accelerated the uptake of digital technologies by everyone from pre-school children to grandparents, from the “mom and pop” shop to women entrepreneurs turning to e-commerce platforms to sell goods and services, as well as the many other occupations that quickly adapted to a new normal for economic survival. With this faster than expected digital transformation, the workforce and especially women will have to adjust and gain new skills to remain competitive and stay employed.

A few years ago, McKinsey estimated that anywhere from 40 million to 160 million women would have to make occupational transitions due to automation. This holds truer today because the pandemic accelerated digitization of industries and services. Globally, policymakers are having discussions on the interplay between skills demand and the role of human capital to foster digital transformations across industries and firms.

In short, countries are reemerging at a time when technology and artificial intelligence (AI) are shaping how we live, work, and play at a faster rate than pre-pandemic levels.  Societies cannot afford to have women lag behind in terms of employment opportunities in both advanced and developing countries. In OECD countries, job growth pre-pandemic had been led by a demand surge for high skills benefitting women instead of men—partly because more women graduate with tertiary degrees than men.

In low- and middle-income countries, women entrepreneurs comprise a large percentage of the labor force. In Africa, women comprise about 50 percent of the continent’s self-employed workforce in the non-agricultural sector. [3] But the impacts of the pandemic stripped away sources of support from both high-skilled and low-skilled women essentially pushing many out of jobs and now preventing them from job seeking—limiting lifetime earnings and stunting a country’s economic growth.  Economic recovery depends on governments ensuring that women have equal access to jobs, digital connectivity, and digital skills. is here to guide public employment services to navigate the post-COVID economic recovery by providing AI-driven digital solutions that empower women and the most vulnerable job seekers in the labor market to match their skills and talents with quality jobs.  Let’s rethink how we approach women’s labor force participation using actionable insights with non-discriminatory and unbiased jobs matching results by contacting and visiting our product site for PES.



[1] Time Magazine. Women are Now the Majority of the U.S. Workforce—But Working Women Still Face Serious Challenges. January 2020.

[2] National Women’s Law Center Factsheet. Ewing-Nelson and Tucker. April 2021.

[3] World Bank. Profiting from Parity: Unlocking the Potential of Women’s Business in Africa. 2019.

“No unemployed candidates will be considered at all” – the crux of unemployment.

Back in 2008, when we first started developing our solutions, the work of Diamond, Mortenson and Pissarides provided the scientific basis for our job and skills matching technology. With their Nobel prize winning labor market theory and the DMP model, they provided a first coherent, complete framework to think about labor market dynamics in a structured way. In their theory, labor markets are viewed as markets with search frictions: workers look for suitable jobs and employers look for suitable workers, both investing considerable time and effort; search frictions are the process, or time factor, of matching the two.

The DMP model itself describes the search activity of the unemployed, the recruiting behavior of businesses and wage formation. When jobseekers and employers find each other, they negotiate wages based on the labor market situation: the number of unemployed workers and the number of vacancies, as well as other factors such as how long it will take to find that job, the workers’ unemployment benefits and what value the worker attributes to not having to work while searching. The model can thus be used to estimate the effects of different labor-market factors on unemployment, the average duration of unemployment, the number of vacancies and real wage. Such factors may include the level of unemployment benefits, the real interest rate, the efficiency of employment agencies, hiring and firing costs, etc.

On-the-job search and its effects on labor market dynamics

This framework significantly furthered understanding of how mismatch problems and a lack of symmetry between different search mechanisms and the resulting imbalance between supply and demand affect the functioning of the labor market. However, one key aspect of the labor market is completely ignored here, namely that not all jobseekers are unemployed. The majority of the literature since then typically also focused on the unemployed, not only because the standard DMP framework does not include on-the-job search, but also due to limited availability of on-the-job search data. More recently, however, research has begun to include on-the-job search and job ladders. The idea of a job ladder is that all workers agree on which jobs are more desirable in the sense of job and wage satisfaction and slowly climb the job ladder from “bad” or unsatisfactory jobs to “good” jobs through job-to-job transitions. Occasionally, negative shocks throw them off the ladder and back into unemployment. A growing number of studies have documented the importance of on-the-job search and its related job ladder dynamics for macroeconomic outcomes.[1] Some argue that the labor market is segmented in that employed and unemployed jobseekers are unlikely to directly compete with each other for jobs because they have different job-relevant characteristics and apply for different jobs. For example, Longhi and Taylor (2013) state that the unemployed only apply for “bad” jobs and the employed for “good” jobs and so they do not compete. However, they do not investigate the reasons for this behavior and it may well be that the cause is somehow tied to the search behavior of employed workers or related dynamics. For instance, they find that a larger proportion of the unemployed “prefer” a part-time job compared to the employed and state that this supports their claim of a segmented labor market, ignoring the fact that this may not be an inherent “preference”, but instead a higher flexibility on part of the unemployed based on their more pressing need to find any employment at all. Even though they note themselves that part-time workers are more likely to search on the job, probably because they are “unsatisfactory in terms of labor supply preferences”. Similarly, they find that the two groups tend to use different search methods, with the employed focusing more on using their networks and the unemployed relying more on job centers and employment agencies. They use this as another argument for their conclusion that they are not applying for the same jobs, apparently because the jobs available through these different channels differ. But this could instead have more to do with the fact that with increasing length of unemployment, jobseekers’ personal and professional networks decline and the unemployed become more reliant on institutional support. It does not necessarily imply that the unemployed actually want to apply for different jobs.

Indeed, the bulk of recent literature finds that on-the-job search has a clear effect on macroeconomic outcomes and the chances of unemployed jobseekers on the labor market. Moscarini and Postel-Vinay (2019) and Faccini and Melosi (2019) link on-the-job search to inflation, arguing that when employment is concentrated at the bottom of the job ladder, typically following a recession, employed workers search to find a better job. As workers climb the job ladder, the labor market tightens and generates inflation pressures through wage negotiations. Eeckhout and Lindenlaub (2019) provide an elegant theory where the search behavior of employed workers generates large labor market fluctuations even in the absence of other shocks through a strategic complementary between on-the-job search and vacancy posting. According to this theory, the labor market itself can generate cycles, contrary to the longstanding assumption (based on the DMP model) that such cycles can only be generated by exogenous shocks. The authors state that active on-the-job search improves the quality of the jobseeker pool, which encourages vacancy posting through firms, which makes on-the-job search more attractive. This corresponds to an economic boom with little mismatch, abundant job creation and low unemployment. On the other hand, during a recession, the jobseeker pool has a much lower proportion of on-the-job searchers. As a result, firms have less incentive to post vacancies, which generates a low matching rate for workers which cannot compensate the cost of on-the-job search, leading to high mismatch and high unemployment. The authors show that their theory, in particular the search behavior of the employed, can explain many important labor market phenomena, including large fluctuations in unemployment and the fact that unemployment rates take much longer to recover than vacancies and productivity, say, following a recession.

It may seem counterintuitive that the behavior of the employed could explain unemployment. But the employed typically have a share of over 90 percent of the labor force and apply for job openings in the same labor market as the unemployed. Therefore, any minor change in their behavior has deep aggregate implications for unemployment. Even if they search much less intensively than the unemployed, on average, almost half of the new jobs are filled by employed workers. Particularly at the end of a recession, the employed searchers crowd out the unemployed ones. As job creation picks up, jobs go disproportionately to the on-the-job searchers and not to the unemployed. All the renewed activity thus initially translates in better jobs for the employed, but not in improved prospects for the unemployed.

Based on a survey that focuses on job search behavior regardless of labor force status, Faberman et al. (2020) find evidence supporting Eeckhout and Lindenlaub’s theory in the following three facts: (1) on-the-job search is pervasive, and is more intense at the lower rungs of the job ladder; (2) the employed are about four times more efficient than the unemployed in job search [2]; and (3) the employed receive higher-quality job offers than the unemployed.

The stigma of unemployment

What these theoretical models and studies do not mention, is why the employed are more successful in job search and receive higher-quality job offers than the unemployed. Much of this may have to do with the stigma of unemployment – especially long-term unemployment [3]. The quote in the title of this article is from a job posting by Sony Ericsson, and they are not alone. Various studies (for example, the ones described here and here or here) have shown consistently over the years that hirers are biased against unemployed applicants, often assuming that the unemployed are lazy, less productive and less competent workers than employed applicants with otherwise equal characteristics. A 2019 study found that, based on stereotypical perception of unemployed applicants, hirers even condemn their character: unemployed job candidates are seen as less warm, less trustworthy, less well-intentioned, less friendly, and less sincere compared to employed job candidates. No wonder the unemployed are forced to settle for “bad jobs” – if they find employment at all.

And this biased perspective is not only found in hirers, it also seems to be widespread among researchers. For instance, at the core of Eeckhout and Lindenlaub’s theory is the implicit assumption that employed jobseekers are more attractive and valuable than unemployed ones (active on-the-job search improves the quality of the jobseeker pool). Even the DMP model takes a stigmatized view of unemployment: the result that higher unemployment benefits raise unemployment rates is rooted in the assumption that higher income through benefits decrease the unemployed worker’s motivation to search for a job and thus to successfully reenter the labor market. To put it bluntly, the model assumes that unemployed workers prefer leisure to work (are lazy) and puts the blame on them (a motivated unemployed person could find a job at any time).

This, together with the fact that research demonstrates that long-term unemployment also leads long-lasting damage such as to lifetime lower wages, increased health issues, lower quality of life and diminished lifespan as well as an increased risk of suicide, clearly shows that unemployed jobseekers should be protected and that efforts should be increased to prevent further unemployment and to mitigate long-term unemployment. One small but simple step is already apparent: promote solutions that prevent this bias, at least in the first steps of the candidate selection process, by using labor intermediation systems that mask labor force status. However, many current systems and platforms offered by PES only provide access to unemployed jobseekers. These systems are rarely successful, often barely frequented by companies and potential employers. And the stigma of unemployment is a key reason for this issue. To be sustainable in the long term and offer unemployed jobseekers a real chance to return to work, a good PES platform must include the whole universe of workers and specialists from all fields and industries and competences.

Of course – contrary to what some software providers claim – simply introducing the right software will neither fully solve the problem of discrimination against the unemployed, nor can it reduce unemployment on its own. This is a complex issue depending on many factors which needs to be tackled from multiple angles. Nevertheless, such solutions can serve as an effective component of well-designed labor market and anti-discrimination policies.

Here at JANZZ, we don’t just go with quick marketing headlines, we develop evidence-based solutions and  have already been deploying them successfully since 2010. Our job and skills matching solutions are fair and non-discriminatory, producing completely unbiased results according to the OECD principles on AI. This guarantees that the best candidate with the best aptitude in all individual criteria achieves the best match – regardless of labor force status or other non-relevant characteristics such as origin, age or gender. Which is one of the many reasons why we are a trusted partner of an ever-growing number of public employment services across the globe.

If you want to take this first step in breaking the cycle and contributing to a fairer labor market for the unemployed, contact us at or visit our product site for PES.


[1] Christensen et al., (2005), Cahuc, Postel-Vinay, and Robin (2006), and Bagger and Lentz (2019), among others,
[2] If they had relied only on transition rates – a common approach in the literature due to lack of data on job search effort – they would have found the opposite result of Fact (2), namely that the unemployed are about seven times more efficient.
[3] In the US, long-term unemployment is defined as (active) unemployment for longer than 6 months; in the EU for longer than 12 months.

AI, automation and the future of work – beyond the usual bubbles

In recent years there have been many posts, articles and reports on how AI and automation will shape the future of work. Depending on the author’s perspective or agenda, these pieces go one of two ways: either the new technology will destroy jobs and have devastating effects on the labor market, or it will create a better, brighter future for everyone by destroying only the boring jobs and generating better, much more interesting ones. As always, the truth probably lies somewhere between these two extremes. In this post, we want to take a more nuanced view by discussing the most common arguments and claims and comparing them with the facts. But before we get into this, let us first clarify what the AI-driven digital transformation is. In a nutshell, it is all about automation, using AI technology to complete tasks that we do not want humans to perform, or that humans cannot perform. Just as we did in the past, in the first, second and third industrial revolutions.

From stocking looms to AI art

With each of these revolutions came the fear that human workers would be rendered obsolete. So why do we want to automate? Even though in some cases, inventors were, and still are, simply interested in the feat of the invention itself, more often than not an invention or development was driven by business interests. As is widespread adoption. And no matter which era, businesses rarely have other goals than staying competitive and raising profits. 16th century stocking looms were invented to increase productivity and lower costs by substituting human labor. Steam-powered machines in 19th century mills and factories and farm machinery were used for the same reason. Robots in vehicle manufacture in the second half of the 20th century – ditto. Whether the technology is tractors, assembly lines, or spreadsheets, the first-order goal was to substitute human musculature by mechanical power, human handiwork by machine-consistency, and slow and error-prone “humanware” by digital calculation. But so far, even though many jobs were lost to automation, others have been created. Massively increased production called for jobs related to increased distribution. With passenger cars displacing horse-powered travel and equestrian occupations, and increasing private mobility, jobs were instead created in the expanding industry of roadside food and accommodation. Increasing computational power used to replace human tasks in offices also led to entirely new products and the gaming industry. And the rising wealth and population growth accompanying such developments led to increased recreational and consumption demands, boosting these sectors and creating jobs – albeit not as many as one may think, as we will see below. However, we cannot simply assume that the current revolution will follow the same pattern and create more jobs and wealth than it will destroy just because this is what happened in the past. Unlike mechanical technology and basic computing, AI technologies not only have the potential to replace cheap laborers, say, with cleaning or agricultural robots. They have also begun outperforming expensive workers such as pathologists diagnosing cancer and other medical professionals diagnosing and treating patients, and are also touching on creative tasks such as choosing scenes for movie trailers or producing digital art. Of course, we should also not simply assume a dystopian future of less jobs and sinking wealth. But we must keep in mind that currently, in many cases it is more cost effective to replace expensive workers with AI solutions than cheap laborers such as textile workers in Bangladesh.

So, working towards a more differentiated view, let us take a look at the currently most common claims and how they stand up to closer scrutiny.

Claim 1. AI will create more/less jobs than it destroys

This is the main argument put forth in utopian/dystopian scenarios, including reports by WEF (97m new jobs vs. 85m displaced jobs across 26 countries by 2025), PwC (“any job losses from automation are likely to be broadly offset in the long run by new jobs created”), Forrester (job losses of 29% by 2030 with only 13% job creation to compensate) and many others. Either way, any net change can pose significant challenges. As BCG states in a recent report on the topic, “the net number of jobs lost or gained is an artificially simple metric” to estimate the impact of digitalization. A net change of zero or even an increase of jobs could cause major asymmetries in the labor market with dramatic talent shortages in some industries or occupations and massive worker surplus and unemployment in others. On the other hand, instead of causing unemployment – or at least underemployment, less jobs could also lead to more job sharing and thus shorter work weeks. Then again, although this may sound good in theory, it also raises additional questions: How will pay and benefits be affected? And who reaps the bulk of monetary rewards? Companies? Workers? The government? It is admittedly too soon to see the effects of AI adoption so far regarding overall employment or wages. But past outcomes, i.e., of previous industrial revolutions, do not guarantee similar outcomes in the future. And even those outcomes show that job and wealth growth were not necessarily as glorious as they are often portrayed. The ratio of employment to working age population has remained fairly constant in OECD countries since 1970, rising only from just over 64% to just under 69%.[1] Much of this increase can be attributed to higher labor participation rates, especially of women. And the increased wealth is clearly not evenly distributed, e.g., in the US:


AI, automation and the future of work – beyond the usual bubbles

Source: Economic Policy Institute,


AI, automation and the future of work – beyond the usual bubbles

Source: Economic Policy Institute,


There are simply no grounds to assume that AI and automation will automatically make us wealthier as a society or that the increased wealth will be distributed evenly. We should thus be equally prepared for more negative scenarios and discuss how to mitigate the consequences. For instance, would it be acceptable to treat AI processes like human labor? If so, we could consider taxing them to support redistribution of wealth or to finance training or benefits and pensions for displaced workers.

In addition, these estimates should be questioned on a basic level. Who can confidently say that this job will decline? How can we know what kind of jobs there will be in future? None of these projections are truly reliable or objective – they are primarily based on some group of people’s opinions. For instance, the WEF’s Future of Jobs Report, one of the most influential reports on this topic, is based on employer surveys. But it is simply naive to think anyone, let alone a cadre of arbitrary business leaders, can have a confident understanding of which jobs and skills will be required in the future. One should not expect more from this than from fortune telling at a fair. Just take a look at the predictions about cars in the early 19th century, remote shopping in the 1960s, cell phones in the 1980s, or computers since the 1940s. So many tech predictions have been so utterly wrong, why should this change now? And yet, such predictions are a key element in the estimates for the “future of work”.

Fact is, scientifically sound research on this topic is extremely scarce. One of the few papers in this area studied the impact of AI on labor markets in the US from 2007 to 2018. The authors (from MIT, Princeton and Boston University) found that greater AI exposure within businesses is associated with lower hiring rates, i.e., that AI adoption has so far been concentrated on substitution as opposed to augmentation of jobs. The same paper also finds no evidence that the large productivity effects of AI will increase hiring. Some people may be tempted to say that this supports the dystopian view. However, we must also note that this study is based on online vacancy data, and thus the results should be treated with caution, as we explained in detail in one of our other posts. In addition, due to the dynamics of technological innovation and adoption, it is almost impossible to extrapolate and project such findings to make robust predictions for future developments.

And on a more philosophical note, what would it mean for human existence if we worked substantially less? Work is ingrained in our very nature; it is a defining trait.

Claim 2: Computers are good at what we find hard and bad at what we find easy

Hard and easy for who? Luckily, we do not all have the same strengths and weaknesses, so clearly, we do not all find the same tasks “easy” and “hard”. This is just yet another extremely generalizing statement based on completely subjective judgement. And if it were true, then most people would probably consider repetitive tasks as typically easy, or at least easier. This directly contradicts the next claim:

Claim 3: AI will (only) destroy repetitive jobs and will generate more interesting, higher-value ones.

The WEF states that AI will automate repetitive tasks like data entry and assembly line manufacturing, “allowing workers to focus on higher-value and higher-touch tasks” with “benefits for both businesses and individuals who will have more time to be creative, strategic, and entrepreneurial.” BCG talks of the “shift from jobs with repetitive tasks in production lines to those in the programming and maintenance of production technology” and how “the removal of mundane, repetitive tasks in legal, accounting, administrative, and similar professions opens the possibility for employees to take on more strategic roles”. The question is, who exactly benefits from this? Not every worker who is able to perform repetitive tasks has the potential to take on strategic, creative and entrepreneurial roles, or program and maintain production technology. It is simply a fact that not everyone can be trained for every role. More satisfying, interesting tasks for intellectuals (such as the advocates of a brighter future of work thanks to AI) may simply be too challenging for the average blue-collar worker whose job – which may well have been perfectly satisfying to them – has just been automated. And not every white-collar worker can or wants to be an entrepreneur or strategist. Also, what exactly does “higher-value” mean? Who benefits from this? The new jobs created so far, like Amazon warehouse workers, or Uber and Postmates drivers, are not exactly paying decent, secured living wages. And since the early 1970s, businesses have demonstrated a clear disinterest in sharing the added value from productivity gains with workers:


AI, automation and the future of work – beyond the usual bubbles

Source: Economic Policy Institute,


On the other hand, a vast number of the AI applications that are already available perform higher- to highly-skilled tasks based on data mining, pattern recognition and data analysis: diagnosis and treatment of medical conditions, customer service chatbots, crop optimization and farming strategies, financial or insurance advising, fraud detection, scheduling and routing in logistics and public transport, market research and behavioral analysis, workforce planning, product design, and much more. The full effect of these applications on the job market is not yet clear, but they are certainly not only removing mundane, repetitive tasks from job profiles.

Claim 4: We (just) need to up-/reskill workers.

While we certainly do not disagree with this statement in general, it is often brought up as a more or less simple remedy to prepare for the future AI-driven shifts in the labor market and “embrace the positive societal benefits of AI” (WEF). Fact is, this comes with several caveats which make it a far from simple solution.

First off, we cannot repeat enough that it is not possible to predict the “future of work” reliably, especially which jobs really will be in demand in future and which will not. Also, based on the effects of the previous industrial revolutions and current research, it is highly likely that widespread adoption of AI will introduce new jobs with profiles that we cannot anticipate at the moment. This means we need to equip current and future professionals with the skills necessary for jobs that we currently know nothing about. One often suggested way to work around this issue is to encourage lifelong learning and promote more adaptable and short-term forms of training and education. This is certainly a valid option and clearly on the rise. However, there are several aspects to keep in mind. For instance, 15–20% of the US and EU adult population[2] have low literacy (PIAAC level 1 or below). This means that they have trouble with tasks such as filling out forms or understanding texts about unfamiliar topics. How can these people be trained to succeed at “more complex and rewarding projects” if they cannot read a textbook, navigate a manual or write a simple report? In addition, around 10% of full-time workers in the US and EU are working poor.[3] These people typically have neither the time, resources, nor support from employers for lifelong learning and thus no well-informed access to efficient, targeted and affordable (re)training.

By the time such issues have been addressed, many of these workers may have already missed the boat. In 2018, US employers estimated that more than a quarter of their workforce would need at least three months of training just to keep pace with the necessary skill requirements of their current roles by 2022.[4] Two years later, that share has more than doubled to over 60%, and the numbers are similar across the globe.[5] In addition, even before the post-Great Recession period, only roughly 6 in 10 displaced US workers were re-employed within 12 months in the 2000 to 2006 period.[6] In 2019, this rate was the same in the EU.[7] With increasingly rapid changes in skills demands, combined with lack of time and/or resources for vulnerable groups such as the working poor and workers with low literacy, not to mention lacking safety nets and targeted measures in underfunded workforce development systems, the prospects for these workers are highly unlikely to improve.

Moreover, the pandemic has massively accelerated the adoption of automation and AI in the workplace in many sectors. Robots, machines and AI systems have been deployed to clean floors, take temperatures or food orders, replace employees in dining halls, toll booths or call centers, patrol empty real estate, increase industrial production of hospital supplies and much more within an extremely short period of time. In the past, new technology was deployed gradually, giving employees time to transition into new roles. This time, employers scrambled to replace workers with machines or software due to sudden lockdown or social distancing orders. This is a crucial difference to the preceding industrial revolutions. Many workers have been cut loose with simply not enough time to retrain. Similarly disruptive events may well occur in the future – be it another pandemic or a technological breakthrough – and as a society, we need to be prepared for these events and provide affected workers with swift, efficient and above all realistic support.

Claim 5: Employers should view up- and reskilling as an investment, not as an expense.

If a company replaces all of its cashiers by robots, why would they want to reskill the newly redundant workers? Even governments have a hard time taking this stance on training and education. Many countries focus primarily on college or other education for young workers rather than retraining job seekers or employees. For instance, the US government spends 0.1% of GDP to help workers navigate job transitions, less than half what it spent 30 years ago – despite the fact that skills demand is changing much faster than it did three decades ago. And the vast majority of businesses is primarily interested in maximizing profits – that is just how our economy works. Remember: we live in a world where even sandwich makers and dogwalkers are forced to sign noncompete agreements to prevent them from getting a raise by threatening to move to a competitor for higher pay.

Well-performing conversational software could enable a company to take a 1,000-person call center and run it with 100 people plus chatbots. A bot can respond to 10,000 queries in an hour, far higher than any realistic volume even the most efficient call-center rep could handle. In addition, a chatbot does not fall ill, need time off work, or ask for perks and benefits. They make consistent, evidence-based decisions and do not steal from or defraud their employers. So, if the quality of this software is sufficient and the price is right, there would most probably be an uproar amongst shareholders if a company did not go for this offer. After all, a solution that increases efficiency and productivity while lowering expenses is the ideal of businesses of our time. So, if this company doesn’t opt for it, its competition will. And despite the “tech for social good” propaganda we constantly hear from Silicon Valley, most companies are simply not interested in the future of soon-to-be-ex workers.

Beyond the bubble

The bottom line is that we cannot afford to overdramatize or simply reassure ourselves that there will be enough jobs to go round, or we will constantly be playing catch up. Most of the commonly cited problems or solutions tend to be discussed within the academic or high-income bubble of researchers, tech entrepreneurs and policy makers, mixed with a substantial amount of idealism. But to get ahead of these developments that – for good or for bad – have vast potential to completely transform our labor markets and society, we need to look beyond our bubble and design realistic strategies for the future based on facts and objective data.


[2]   US:
[3]   US:
[4]   The Future of Jobs Report 2018, World Economic Forum, 2018.
[5]   The Future of Jobs Report 2020, World Economic Forum, 2020.
[6]   Back to Work: United States: Improving the Re-employment Prospects of Displaced Workers, OECD, 2016.


Building the AI-ready workforce: China’s Artificial Intelligence Plan pushed by both central and local governments

This is part of a series of articles we conduct to analyze government policies and practices on the strategies to build AI workforce. Previously, we have analyzed how Singapore is helping mid-career PMETs to switch to the tech sector and a collaborative effort between government, tech companies and education providers in Saudi Arabia. Our third stop is China.

As the world’s major economies have announced the development of artificial intelligence as a national strategy, China as well released its New Generation Artificial Intelligence Development Plan in 2017.  It is the first time AI was specifically mentioned in the country’s national work report and China aims to build its core AI industry worth over 400 billion RMB (around 57 billion USD) by 2025 and to become the world leading AI power by 2030.

Chinese Ministry of Industry and Information Technology has recently published a report on AI talent development. The report identified three major issues in its current AI talent pool, which are a significant mismatch between skills and jobs, a short supply of highly skilled AI talents, and a regional unbalance of AI talents. It has forecasted a shortage of 300,000 AI workforces in the coming years. To tackle these problems, joint efforts from central and local governments have been made to actively promote the building of AI talents.

In April 2018, an action plan has been laid out for higher education institutions by the country’s Ministry of Education. The plan urged to incorporate AI programs into higher education curricula, to build 100 “AI+X” models to nurture AI talents in specific fields by 2020, to build 50 AI colleges, AI research centers or cross-collaborative research centers by 2020 and to introduce AI in primary and secondary schools. According to the announcement on their website, a total number of 215 higher education institutions currently offer AI major for undergraduate students.

In January 2020, the Ministry of Education, along with the National Development and Reform Commission and the Ministry of Finance announced a joint circular to further develop the “AI+X” model and promote postgraduate students in the AI fields. During the same year, the Ministry of Education expanded the number of master’s students by 189,000 and AI is among the popular majors for expansion [1]. Proposals in the joint circular include encouraging corporation of leading AI enterprises and universities, offering flexible employment models to attract AI experts from enterprises and research institutes to work in universities and funding projects involving collaborative training of AI students between enterprises and universities.

The local governments have as well launched their efforts. The Beijing-Tianjin-Hebei region, the Yangtze River Delta region, the Guangdong-Hong Kong-Macau-Great Bay Area and the Sichuan-Chongqing region are the main development heights of the AI industry in China. Chongqing, China’s industrial heartland city has become one of the 13 “AI pilot zones” in China in early 2020. As a means of China’s AI strategy, such AI pilot zones are going to be granted with financial support and favorable local regulations to encourage the expansion of AI industry. According to Chongqing government website, the first 73 major projects with a total investment of about 29.6 billion RMB (around 4.2 billion USD) have been approved.

In the meantime, the government of Chongqing has released several policy measures to expand the city’s AI talent pool, including releasing an in-demand talent barometer to perform statistically analyses and forecast on workforce, financially supporting AI enterprises to carry out one-to-two-year apprenticeship in AI related positions, and financially supporting AI enterprises to offer AI related internship positions with no more than 12 months. was invited twice to Chongqing in 2019 thanks to the Consulate General of Switzerland in Chengdu and Sino-Swiss TechnoPark, and we were also horned to participate in the Smart China Expo a high standard tech event hosted annually in Chongqing. We have witnessed the huge potential within this traditional manufacturing city. As worldwide governments embark on the journey of digital transformation, their public employment services (PES) seek suitable solutions to support them in the increasingly important role they play in job matching, enhancing employability, addressing skill gaps and aligning education offerings with market needs.

To learn about how assists PES in tackling these challenges, please visit our product site for PES or contact us at