It seems that every week, AI technology has learned to do something humans do, but faster and better. From detecting cancers and eye conditions to predicting floods; or analyzing the language, tone, and facial expressions of candidates during recruitment processes, AI is now at the stage where it not only supports human judgment, but also makes increasingly more complex and accurate decisions.

As technology further improves and we learn how to better work and collaborate with AI, interactions between humans and computers will significantly enhance creativity—of both humans and bots.

Technology is supposed to help us, but sometimes it feels like for every step forward, we take two steps back. Like many people (and despite my resistance), my family has accumulated a few internet of things (IoT) devices in our home. Our motivation for acquiring them has been to streamline our lives.

For instance, when we needed to be able to play our newborn son’s music anywhere in the house, a Bluetooth speaker seemed a natural fit—and it was. Inspired, we adopted a “voice assistant” for the kitchen to put all the information and entertainment of the internet at our fingertips while we prepared meals. This works great most of the time. The rest of the time, the voice assistant (who shall remain nameless) just ignores me! And if my wife tries, good luck. She’s ignored more often than not; at least that’s how it seems.

Issues with functionality aside, I was resistant to the haphazard introduction of IoT devices into our home for other reasons. Because I have worked closely with computers and technology for my whole career, the fact that these devices are computers is not lost on me. I also realize the many implications of putting an internet-connected, microphone-equipped computer on my kitchen counter. (Will a live feed from my house wind up on a sketchy website?)

Imagine that your boss Ethan calls you into his office. He expresses disappointment in your recent performance and lack of commitment. How would you react? Accept the feedback and put in more effort? Would you pout in your office and start looking for a new job?

Now, would your reaction be different if your boss was not named Ethan but Emily?

I’m a professor of economics, and my research investigates this very question. We hired 2,700 workers online to transcribe receipts, randomly assigning a male or female name to a manager, and randomly assigning which workers would receive performance feedback.

Results show that both women and men react more negatively to criticism if it comes from a woman. Our subjects reported that criticism by a woman led to a larger reduction in job satisfaction than criticism by a man. Employees were also doubly disinterested in working for the firm in the future if they had been criticized by a female boss.

Technology companies are frequently driven by their engineering processes. Of course product quality is regarded as most important, and that quality can be tested and measured with numbers and data. Such companies also frequently align their core identity with the engineering that belies their innovation. Their top executives often started out as engineers and keep looking primarily through their engineering lens as they become company leaders. Although it makes perfect sense, this approach is misguided.

In the past two months we have looked at how three-sigma limits work with skewed data. This column finds the power functions for the probability limits of phase two charts with skewed probability models, and compares the trade-offs made by three-sigma limits with the trade-offs made by the probability limits.

Phase two charts

Ever since 1935, there have been two approaches to finding limits for process behavior charts. There is Walter Shewhart’s approach using fixed-width limits, and there is Egon Pearson’s fixed-coverage approach based on probability models. (For more on these two schools of thought, see “The Normality Myth,” Quality Digest, Sept. 19, 2019.) About the year 2000, some of my fellow statisticians tried to reconcile these two approaches by talking about “phase one and phase two control charts.”

Phase one charts use Shewhart’s fixed-width, three-sigma limits. These charts are used to help identify assignable causes of exceptional variation so that the process can be adjusted or fixed as needed. Then, under the assumption that once a process is fixed it will stay fixed, it is time for phase two.