Hello there!
In the future, AI will be present in most areas of our lives. There won’t be a way around it: Getting a job, keeping it, or transitioning to a new role will likely be in some way related to AI. But this doesn’t mean that it’s a bad development.
In this issue, I collected stories that show you how to be mindful about algorithms in recruiting and how you might benefit in the future from AI job recommendations.
Collecting the stories for this issue, it again became clear that there is a lot of talk about automation, AI and the effect of algorithms on work, but there is no need to panic.
I’m curious to read what you think about those stories. Feel free to drop a comment!
Best,
Alice
Headlines you shouldn’t miss
ZDNet Online workers are creating new tools to improve their working conditions. Will it work? Gig workers at Amazon Mechanical Turk have been exchanging ideas and experiences on online forums and websites about the nature of gig work and opportunities to gain more money. Yet, they are far from establishing classic bargaining power.
WONDERFUL ENGINEERING This Cafe In Japan Has Robot Waiters That Are Controlled By Disabled Workers: At the Dawn Avatar Robot Café in Tokyo, robots are placing and delivering orders. Those machines are not highly advanced independent waiters but controlled by disabled workers. They can move and communicate through the robot, assuring the right service.
BBC Why artificial intelligence is being used to write adverts: Artificial intelligence can filter, analyze and create new advertising slogans. Using scientific rigor, some ad agencies turned to technology for creative input. While this approach clashes with the old ways of creating ads, AI-driven campaigns can find the right message to cut through the noise of infinite stimuli in the modern world.
POLITICO How the pandemic ate millions of jobs in American restaurants: Covid-19 has disrupted the leisure industry in an unprecedented way. Millions of waiters lost their jobs while consumers shifted to online delivery. Even with the economic recovery, restaurants are not hiring waiters at pre-pandemic rates. However, some have embraced technological change and changed the roles of workers to cater to digital customers.
FORBES A New Kind Of Alien: AI And The Hiring Process: Applicant Tracking Systems (ATS) make recruiting cheaper, more efficient, and improve compliance. However, these algorithm-based systems favor simple formatting. Applicants must adjust their CVs to ATS as they are increasingly popular in HR.
AI can help you transition careers
Globalization, digital advancements, and crises like the Covid-19 pandemic have forced people to become more flexible about their jobs. Nowadays, people in Western societies change their careers more often than in the post-war world. In Australia, researchers suggest that workers will have seven different careers in their lifetime.
Automation will perform certain tasks in the future, while new jobs will emerge. But if you lose your job, how can you know which route to take next?
Nik Dawson, Honorary Scholar (University of Technology Sydney), Marian-Andrei Rizoiu, Lecturer in Computer Science (University of Technology Sydney), and Mary-Anne Williams, Michael J Crouch Chair in Innovation (UNSW), have developed a system that maps skills, the likelihood of automation and the probabilities to transition into related jobs.
The researchers analyzed 500 skills in 13 clusters (e.g., healthcare, software development). Based on this skill assessment, they could see which jobs required similar abilities.
They added a layer of automation risk to the data and could see which jobs have the highest risk of automation. Now they knew which jobs were threatened by automation and which skills workers could use from their experience to transition to another job with similar requirements. They developed a transition map to see between which occupations transition would be likely:
The AI-driven system is one example that shows how data can help people find the best fit for a career change. Regardless of whether workers decide to change careers due to unemployment or seek a new challenge, finding the right fit is a crucial component for a successful job transition.
How to get your CV AI-proof
Imagine you spot the perfect job opening for you. It’s a challenging position, the tasks look demanding, you’re perfectly qualified and the salary is above average. You decide to give it a shot and update your CV in an attempt to stand out. Conventional wisdom tells you to be mindful of design choices and the description of your previous experience. So, you craft a creative CV, send the application — and nothing happens.
What might have happened is that your CV was filtered by software searching for the keywords used in the opening. And as you decided to test a creative approach, your use of language and design didn’t correspond with the algorithm’s settings and you’re application got sorted out before a human recruiter could see it.
“You want the simplest, most boring résumé template you can find. You want to write like a caveman in the shortest, crispest words you can,”
says Ian Siegel, Co-Founder of ZipRecruiter. He’s an expert on algorithm-based recruiting, and in an MIT Technology Review article, he recommends the following rules of thumb:
Use short, descriptive sentences
Clearly list your skills
Include details about where you learned your skills
Add licensing and certification numbers if possible
“You want to be declarative and quantitative, because software is trying to figure out who you are and decide whether you will be put in front of a human.”
Before you upload your CV, consider testing it online: Jobscan and VMock allow users to get their CVs analyzed by an algorithm and see if they would pass the first round.
The robots aren’t coming — at least not to US SMEs
The fear of automation concerns many workers. In combination with modern robotics, AI delivers a powerful combination that could easily replace workers and boost productivity. However, small and medium-sized enterprises (SMEs) are cautious and not purchasing the new productivity boosters.
In many Western countries, SMEs are the backbone of the economy and employ a significant share of the workforce. Researchers have expected that those companies would purchase new robots to compete on a global scale.
However, it doesn’t seem to be the case. Researchers Suzanne Berger and Anna Waldmann-Brown of the Massachusetts Institute of Technology (MIT) have looked at 34 SMEs in Ohio, Arizona and Massachusetts. They found that only one had purchased a significant number of robots in the last five years. This particular company was owned by a Japanese corporation financing the investments.
Lack of cash and few long-term prospects seem to be the main drivers. For SMEs, investing in robots requires security beyond a few months, which many don’t have.
The MIT researchers express concerns about the inability and reluctance to invest in robots. SMEs could become less competitive over time, risking getting out of business.
Are AI economists solve the biggest questions of wealth distribution?
Tohid Atashbar, an economist at the International Monetary Fund (IMF), predicts that AI will gradually substitute economists in the future. The newest generation of AI models can learn unsupervised and process a massive amount of data relatively quickly. This means that machines will detect and understand relations of wealth, poverty, tax, income and prosperity, which might be too complex for humans.
Today, AI economists do not exist yet, but algorithms can express the status quo of the economic consensus. The language processing algorithm OpenAI’sGPT-3 is considered to be the most advanced one at the moment. Atashbar has used it to ask questions economists are currently debating:
Question: Is wealth tax effective in reducing inequality?
AI: No, it is not. The most important reason for this is that high income earners are much more mobile than the majority of the population. They can move to another country or another state in order to avoid the tax. If you want to tax the rich, you have to have a global tax.
While this answer looks fairly sophisticated, Atashbar argues that the algorithm seems to have a bias (in this case, pro global tax). He explains that currently, bias is one of the biggest issues as the type of data algorithms use to learn and the methodological preferences can skew answers. However, the economist believes that there is a fair chance of better data leading to better results. Atashbar believes that more neutral AI policy recommendations can diminish the differences between the leading schools of thought in economics (state-friendly vs. market-liberal) and propose more practical solutions.
Opinion: Using AI to assess personality types might be entertaining but misleading
I read an interesting Entrepreneur piece by Martin Rowinski, CEO of Boardsi. He argues that recruiters should turn to AI to analyze candidates’ personality types and find the best-fitting person for a position. In his view, personalities, interests and motivations are crucial aspects of the workplace. Workers need to feel comfortable with the tasks, the expectations related to the job, and the growth prospects. Rowinski sees AI-driven software as a useful tool to be sure if a candidate really fits.
In my personal opinion, it’s too early to trust AI with recruiting to the degree that it assesses candidates’ personality types. In another Jobs Meet Tech issue, I wrote about the shortcomings of recruitment AI. It’s often superficial, flawed and misleading. Sometimes algorithms even confuse the language a candidate is speaking. Based on the biases and immaturity of AI-driven recruiting software, it is hard to imagine that a system could reliably assess something as complex as the human personality.
First of all, candidates tend to present themselves differently when applying for a job — you see one side of an applicant and not their natural behavior. Second, people should be allowed to grow and evolve when commencing a position — sometimes, you only carve out your own potential when applying yourself. The best workers might need time to open up to new tasks.
Most importantly: If recruiters are not aware of the existing shortcomings of AI, they might blindly rely on a system that selects candidates fitting a certain norm but aren’t necessarily the best fit. What do you think? Should recruiters trust this type of AI system?
Tweet of the week: Listen to Elisabeth Reynolds on the future of work
MIT scholar Elisabeth Reynolds explains that technology introduces new types of work, but it is yet unclear how the quality of this new work will look like.