mwhittenberger

  • Section J
  • Section 1
  • Alumni

Submitted

Humans versus robots: When we take performance tracking too far
mwhittenberger
Last modified on April 13, 2020 at 12:55 pm
Amazon has experimented with a variety of tracking systems for warehouse employees, including tracking wristbands and automated termination. What are the human costs of these systems, and how can they be best corrected?

Activity Feed

On April 13, 2020, mwhittenberger commented on People analytics during a pandemic: new opportunities? :

Thanks for a thoughtful post! Though it’s certainly tempting from a data collection perspective, I agree that now is not the time to roll out these kinds of tools and would echo the reasons you highlighted as the primary reasons why. I would also add that ramping up quickly on large amounts of data may actually be bad for the companies that make these products, as well as for the companies that use them and their employees. Scaling issues aside, as you said, these are not “normal times” and their models are theoretically built for such. Will the insights that the tools yield hold any weight, now or even after the pandemic? This could be damaging to customer relationships and the industry overall. Could the makers of the products train algorithms based on massive amounts of work-from-home data only to find that the models are not predictive when everyone goes back to work? As for the users, even with the real-time feedback tools for video calls that are designed for immediate improvements, I question whether those individuals who already struggle with Zoom would become even more overwhelmed, causing them to interrupt or raise their voice even more!

On April 13, 2020, mwhittenberger commented on Forever a Student Driver? Data Monitoring in the Trucking Industry :

Thanks for an interesting post, John! I agree with the points above SmartCap controlling the data and working with various stakeholders to determine the best way balance effective data use with privacy. In addition, I think that SmartCap could work with employers and drivers to use the data even more proactively. They could use anonymized SmartCap data across drivers (and even companies) along with other data sources to identify patterns that may explain or predict driver fatigue before it happens. For example, are there certain routes that are take longer than planned? Is training or intervention x useful in reducing the number of drivers that stretch beyond their limits? Can we nudge this driver to consider whether they are too tired to make this journey? Though the misaligned incentives will persist, and are certainly strong counterforces, these kinds of predictions may help both companies and drivers make safer decisions.

On April 13, 2020, mwhittenberger commented on Books and Movies in the era of AI :

Thanks for an interesting post (and some laughs), Haerin! I agree with @Miho’s point about the story and emotion behind an artist or writer and the impact these elements have on my experience as a consumer of their work. This may be less true for trashy series on Netflix or one-hit wonders, but I feel a human connection to my favorite writers and musicians despite never meeting them. That may explain why the writers and performing artists who are most successful in the long-term build a following around who they are as people, beyond just their writing or music. Algorithms can’t do that, even if they can optimize music to current tastes. [If anyone wants to try their hand at AI-generated music, check out: https://magenta.tensorflow.org

User deleted this comment on April 13, 2020
On April 13, 2020, mwhittenberger commented on Books and Movies in the era of AI :

Thanks for an interesting post (and some laughs), Haerin! I agree with @Miho’s point about the story and emotion behind an artist or writer and the impact these elements have on my experience as a consumer of their work. This may be less true for trashy series on Netflix or one-hit wonders, but I feel a human connection to my favorite writers and musicians despite never meeting them. That may explain why the writers and performing artists who are most successful in the long-term build a following around who they are as people, beyond just their writing or music. Algorithms can’t do that, even if they can optimize music to current tastes. [If anyone wants to try their hand at AI-generated music, check out: https://magenta.tensorflow.org/%5D