top of page

Return to Slavery (by Algorithm)

We Work to Serve Algorithms

“The difference between technology and slavery is that slaves are fully aware that they are not free” — Nassim Nicholas Taleb

I recently read a thread of responses to a post in FB about Lyft and the inequities suffered by the drivers. There are more stories hitting the street (pun intended) about mistreatment of workers at the hands of … algorithms. Lift and Uber are showing us the possibilities of AI being driven by market forces without any social or moral influences to balance the judgement of the algorithms. Markets are designed to extract the most wealth from consumers and maximize return on investment by minimizing the costs to produce the goods and services. In most organizations, the cost of goods sold (CGS) is human labor (combination of salary + benefits). The “good news” for many employers is the trend towards offsetting the costs of benefits by shifting the burden to the employee (either through direct transition of the carrying costs or through elimination via limited hours — partial employment).

According to the Bureau of Labor Statistics, there are about 155 million workers in the US (seems a bit low, but probably there are many who work for cash and are not accounted for…). The number of folks who are working part time (by choice or otherwise) is about 22% and the rest 8% appear to be not working (by choice or otherwise). Even if these numbers are not fully representative of the population of working individuals, the numbers do not capture the challenges faced by many who are fully employed since wages have stagnated for the last 30–40 years (ref. Pew report).

If market forces are left to themselves, then they will naturally direct the use of technology to meet their goals (stated above). The increasing investment in AI by organizations will accelerate the realization for maximizing ROI (return on investment). The examples of Lyft and Uber provide us a glimpse of what (possibly) is a “working” model for many organizations. Initially, algorithms will learn how to do the work (trained by employees) and then they will replace these workers. The replacement will be uneven since there are real limitations to the capability of automation (physical and cognitive) at the moment. But, the use of algorithms to optimize profit has been in place for many years. Visit the stock exchanges and look how few people are involved in the actual trading on the floor. Most of the workers have been displaced by algorithms.

A brief study of the business models underlying Uber and Lyft, reveals the true value in these companies are in the algorithms. This fact is not lost on the workers are are working for wages that are not sustainable (some have already labeled these as slave wages). Unfortunately, this means that the drivers are in fact slaves to these algorithms. What we are seeing is how well these algorithms can extract value from the market (consumer’s wallet) as well as the employee (wage reduction). Just imagine how algorithms driven by market forces can influence other business models.

So, what are the options for balancing the inequities of market forces ? The old standby is for the government to force the hand of the employer to pay a fair wage. But the minimum wage only works if the calculation is based on a part time shift (4hrs) or full time shift (8hr). But, what if the actual productivity is measured by units, where the actual unit of work is (in the case of an Uber or Lyft driver) the transport time from pickup to destination. That means, the lag time between calls, and finding passengers, and the transaction time for collecting fees are not necessarily included in the calc for unit productivity. So, the actual hours of work is far less. This works fine for a machine, but demoralizes and enslaves the human to work many more hours to achieve a decent living wage.

You can say “well, I’m a biochemist working in a lab..that’s never going to happen to me”. Maybe..but what if unit productivity for a researcher was measured by the actual bench time testing and producing new product ? Or a developer who is compensated based on the number of keystrokes per day (as a measure of productivity). Or a manager who is measured by the number of decisions they make vs the number of meetings they attend (hmm..yikes).

So, yes many folks are (rightfully) concerned about the next wave of technology and how it may effect their own revenue (wages) and eventual employment (available jobs not performed by automation). When you think about the meteoric rise in costs for education (public schools transitioning to charter schools) and higher education (college & graduate studies), the burden of debt needs to be balanced by the promise of good paying jobs. It’s no surprise that the levels of anxiety are increasing and people are looking to politicians for answers (any politician) since the realization that our market driven economy will not go quietly into the night.

There may be another option to achieve a better balance for workers and perhaps change the approach for how we use and deploy technology. If technology is used to support human activity, it does not necessitate the replacement of humans in the workforce. What if companies like Uber & Lyft were to shift ownership to the employees ? The co-op model has been around for many years and is seeing increasing traction in other countries and in the US. Most of the co-ops are smaller transitions, from family owned to worker owned organizations. Think about this as a way to democratize work through shared ownership. If we applied this ownership model to Uber & Lyft, the new owners can modify these algorithms to address the inequities so everyone can make a fair living. The ownership transition can be introduced through a number of options, but let’s focus on some emerging (more positive) ones.

It may also be possible for a pooled fund to underwrite an ownership transition, but these are available for smaller organizations.Another option would be for a local startup to duplicate the business model using their a different set of algorithms that provide for an equal distribution of wealth and open it up to Uber & Lyft drivers who will likely step away from their current employers (in a heartbeat, once the new business model proves it can scale).

Those are just a few possibilities for democratizing work, ensuring workers are treated and compensated for their value (instead of paying for their unit cost plus incremental compensation), and restoring the promise of work and value for human potential.

Perhaps we can move away from the trend towards modern day slavery and introduce a morally balanced market approach where we don’t have to treat humans as units of productivity.



bottom of page