The 5 Best Podcasts on Algorithms and Work

Interested in how algorithmic systems affect us at work? Here are some well-researched podcast episodes to get drawn into.


24 October 2023

#ai #podcast #work

Foto von FPVmat A auf Unsplash

I find podcasts a great format to immerse myself in new topics and find out about scientists working on cutting-edge projects while doing other stuff. But there are so many out there that it can be hard to find the really good ones. To help you not get lost in the weeds, I’d like to share my favorite podcast episodes focusing on algorithms and work.

No. 1: Gigaverse by Radiolab

🎧 Listen here

What it is about: Three stories about people’s personal experiences with platform work in the United States

Why I liked it: The second story about a Lyft driver shows really well how algorithmic management common in the gig economy manipulates workers. Over the years, he tried to figure out how to make the most money by playing the “game” - but every time he uncovered new loopholes, they would be closed just as fast. His story also outlines the extreme psychological effects that working for a delivery app can have. The instant rewards and incentive schemes make it so addictive that despite having a great deal of flexibility on the job, it is very difficult to resist going out and deliver when the app’s algorithm is pushing you to do so.

The third story is about people working as shoppers for an app called Shipt. When many of them realized that their earnings per task had become consistently lower or consistently higher all of a sudden, they tried to find out why. They started collecting data, basically using citizen science to get to the bottom of how the payment logic of the app worked. This showed how far the platform would go to avoid being transparent about their algorithmic systems. But when people don’t know how their earnings are calculated, they also can’t figure out how to make more money from the app and whether working for the app is even worth it.

No. 2: Blood in the Machine by 99% Invisible

🎧 Listen here

What it is about: The history of the Luddites 

Why I liked it: The parallels between the struggles of workers against company owners who want to introduce automation in the early 17th century and the situation of workers now is astonishing. When workers were prohibited from unionizing and fighting for better working conditions, the Luddite movement led riots and smashed the automated weaving machines. The uprising was violently put to a halt. Back then, a narrative was created that still persists today, the narrative that anyone who questions the use of technology in the workplace is against progress. Brian Merchant, the journalist who made this episode, just released a book on the same topic.

Bonus: There is also a lovely comic by Tom Humberstone explaining the history of the Luddites and their relevance for today.

No. 3: Chatbots Won’t Take Many Jobs by Tech Won’t Save Us

🎧 Listen here

What it is about: An exploration of the effects of algorithmic systems on jobs

Why I liked it: Host Paris Marx and researcher Aaron Benanav take apart why many of the predictions around the effects of automation and Artificial Intelligence on the labor market should be taken with a pinch of salt. Firstly, asking computer experts what jobs will be automated in the future might not be a conclusive approach, as they don't have any expertise in the respective job profiles and don’t know skills the jobs require. Secondly, there tends to be an overestimation of what technology will be able to do in the future - the automation gap usually takes longer to close than anticipated. Very few jobs are completely automated away, most of them change over time. Thirdly, technology is not the only factor that influences automation - politics, regulation, and economic factors also come into play. Marx and Benanav also remind us that what to think about these technologies depends on what we want to achieve on a societal level by using them.

Bonus: While the episode with Aaron Benanav focused more on not overestimating the effects of Artificial Intelligence on jobs, Paris Marx also talked to artist Molly Crabapple about Why AI is a Threat to Artists, highlighting the issues that already come up around the use of Generative AI systems.

No. 4: Hired by an algorithm by In Machines We Trust

🎧 Listen here

What it is about: The use of algorithmic systems in the recruitment process in the United States

Why I liked it: Hilke Schellmann provides thorough reporting on what algorithmic systems are currently in use in the United States. By providing many examples, the issues coming with them become clear, such as the systems' mistake to treat correlations in data as indicators for certain job candidates' success. The use of behavioral data in algorithmic systems can lead to biases. On LinkedIn, for example, men are more willing to indicate their skills than women. This behavioral difference leads to a disparity in what job ads are being shown to men and women. What came as a surprise to me: Companies do not have to provide proof that the systems they are selling actually work in the way they claim.

Bonus: This is the first episode in a four-part series. The other episodes are also recommendable: Want a job? The AI will see you now, Playing the job market, and Beating the AI hiring machines.

No. 5: Does Artificial Intelligence threaten decent work? by The Future of Work Podcast

🎧 Listen here

What it is about: The effects of Artificial Intelligence on the quality of work

Why I liked it: Hearing about what effects Artificial Intelligence can have on jobs often remains in the realm of predictions or hypothetical assumptions. This ILO podcast with Virginia Doellgast illustrates and spells out what the effects actually are. Firstly, there is labor replacing which means jobs or tasks are automated. Secondly, there is labor controlling, or algorithmic management, which means that workers are controlled and managed by an automated system. Thirdly, there is labor displacing which means the job location changes because of remote work. An increased productivity through algorithmic systems in the workplace can lead to a higher work intensity which is one of the reasons why the workers' input is crucial to make these tools work well.