Machine Learning Tech

Machine Learning Could Check If You’re Social Distancing Properly at Work

Machine Learning Could Check If You're Social Distancing Properly at Work

Recently, Andrew Ng’s startup Landing AI developed an exclusively new workstation monitoring tool that usually provides an alert whenever anybody is less than the desired distance from the other colleague.

Six Feet Apart

A few weeks ago, the startup released a recent blog post with a new demo video that is exhibiting a new social distancing detector. On the left is a feed having lots of people working within the perimeter of the street. While on the right, the bird’s diagram portrays everyone as a dot and then turns them bright red anytime they are moving close to another person.

The company made a statement that the tool is intended to be used in work settings such as factory floors etc. And was created based on the request of their customers (including Foxconn). They also said that the tool could be integrated into an already existing security camera system without experiencing any issue, but that is still systematically examining or investigating how to inform people whenever they violate the social distancing rule.

The company also stated that another possible approach is an alarm that will sound whenever the workers are walking too close. A report can also be generated throughout the night to assist managers in rearranging the workstation. 

Under the Hood

The detector must be calibrated initially in other to map any security footage against the dimensions of the real-world. A well-trained neural network then selects the people in the video, and another algorithm calculates the distances between those people.

Workplace Surveillance

This concept isn’t a recent one. Earlier this year, a report from Reuters said that Amazon is also making use of the same software to monitor the distances between the warehouse team members. The tool also connects an increasing suite of technologies that companies are using intensely to keep their workers under surveillance. 

Currently, there are multiple, cheap, commonly-affordable AI systems that enterprises can buy to watch and monitor all the employees in a store or listen to all the customer service representatives on a phone call. Such as Landing AI’s detector, these systems usually give signal warnings in real-time when behaviors swerve from a specific standard. This pandemic caused by coronavirus has only quicken this trend.

Dicey Territory

The Landing AI stresses that the tool is intended to ensure that it keeps the employees and communities protected and safe, and it should be used with transparency and solely with informed permission. But the same technology can also be misused or used to standardize lots of dangerous surveillance measures.

When inspecting the increasing use of workplace surveillance in its yearly report last December, the AI Now research institute also emphasized that in almost all cases, workers have little power to battle such technologies.

The use of these systems, it wrote, pools power and control in the hands of employers and harms mainly low-wage workers (who are disproportionately people of color). Put another way; it makes an existing power imbalance even worse.

It was also written in the report that the use of these systems puts the total power and control in the care of the employers and majorly damages the low-wage workers (Who are majorly out of proportion people of color). In another approach, it makes an already existing power imbalance extremely bad.

Related posts