Hickok: AI monitoring of workers has ethical, legal and accuracy issues

In this new age of remote working, businesses have unleashed computer monitoring software to monitor workers’ productivity. This “bossware” tracks tasks like typing, video calls, web browsing, and emails. It can even be used to remotely activate microphones and webcams. 

These software tools claim to help companies monitor employees, quantifying their productivity while on the clock. Increasingly, AI is being incorporated into monitoring software to make sense of the large amounts of data being collected. Sometimes AI is also used to monitor employees using facial recognition or object detection technology. 

The Guardian and the Society for Human Resource Management (SHRM) both spoke to Merve Hickock, lecturer at University of Michigan School of Information and founder of AIeithicist.org about the problems and limitations of AI monitoring in the workplace.

Hickok says these AI monitoring technologies have issues when it comes to making judgements. AI models can be trained on data that is inaccurate and biased, leading to inaccurate conclusions about race and gender. 

She explains that remote monitoring that uses a webcam can be particularly problematic. “It gives employers a different level of information than they would have otherwise,” says Hickok.  This sort of intrusion allows employers to see into the private lives of employees, including things like family life or sexual identity.

AI monitoring systems also make judgements on idle vs productive time, classifying what a worker does during the day, says Hickock. She explains that if a worker steps away from their computer to help train a colleague or take a phone call about a project, that time could be considered “idle.” 

Hickok says companies often implement bossware without a full picture of technology blindspots. "A lot of companies don't know what they don't know—they don't necessarily understand when they are bringing AI into the organization," says Hickok. "Some have an inkling about potential bias and some of the risks but may not know what questions to ask or where to begin."

She says that companies should fully understand the limitations, privacy concerns, and ethics of monitoring software before launching. Companies should also audit the software for biases, as well as frequently reviewing new civil rights laws and regulations.

RELATED:

Read “‘Bossware is coming for almost every worker’: the software you might not realize is watching you” on theguardian.com.

Read and listen to “Regulations Ahead on AI” on shrm.org.

Learn more about lecturer Merve Hickok.