[TechTalk] Robots Are Coming to Walmart. Are The Hackers Far Behind?

Aren’t robots just another endpoint attack surface?

About a month ago I saw my first cute robot in a supermarket. Store employees told me it scouts for slippery stuff on the floor and then calls in the real people for the clean up. It looks like the robot in the WSJ story.

Today’s Wall Street Journal reported that Walmart will be expanding its use of robots to take inventory, clean floors and unload trucks ostensibly so they can free up employees to do other tasks.


Walmart robots and others like them are being used to control costs and they say they will free up time for employees to do other work the future. I hope thats so.

Robots will continue to take over more tasks. That is already happening in some industries and it’s a certainty that it will become more prevalent even in the medical professions such as radiology. Machine learning and AI algorithms can easily adapted to look for patterns in scans. It can also be used to do the reverse (create deep fakes and someone just did a proof of concept on a deep fake with software that either put in or took out tumors in a CAT scan. That’s for another post).

According to the WSJ, “The country’s largest private employer said at least 300 stores this year will add machines that scan shelves for out-of-stock products. Autonomous floor scrubbers will be deployed in 1,500 stores to help speed up cleaning, after a test in hundreds of stores last year. And the number of conveyor belts that automatically scan and sort products as they come off trucks will more than double, to 1,200. The company said the addition of a single machine can cut a few hours a day of work previously done by a human, or allow Walmart to allocate fewer people to complete a task, a large saving when spread around 4,600 U.S. stores. Executives said they are focused on giving workers more time to do other tasks, and on hiring in growing areas like e-commerce.”

The company that writes the software for the Robots, Brain Corp said:
“It’s very hard for employers to get the workforce they need,” Mr. Duffy said. “None of the customers we’re working with are using our machines to reduce their labor costs; they’re using them to allow their teams, their janitorial teams, to perform higher-value tasks.”

These are the only the first wave of robots designed to do simple or repetitive tasks.

AI assisted robots eventually will have the ability to displace even highly technical professionals. How fast that will happen depends on many of the unintended consequences and speed bumps along the way of which security is almost a given. Autonomous cars are a good example of this as the date keeps getting pushed further out for their ubiquitous rollout. It was only a year ago that some pundits were predicting autonomous cars would erase the human driven cars by 2022. I don’t see that one happening that fast. A Tesla was recently hacked by researchers and drove onto the wrong side of the highway.

Endpoint security is already a huge problem. We should expect the bad guys to see robots as a target rich surface area once they are more fully deployed and tackling mission critical functions.

Do you think robots represent a serious future security challenge? Bruce Schneier’s new book flags endpoints as potentially game-changing security risks particularly when they control critical functions that could affect safety. Click Here to Kill Everybody -Security and Survival in a Hyper-connected World


Good article that provides some real world perspective behind the reality of Robot job stealing. There are parts of jobs that Robots cannot easily poach from humans because we are better. Radiology is often cited as the one medical profession that can be replaced but as the article in Wired points out –

"Let’s take one example. There are 27 distinct tasks that a radiologist does. One of them is reading medical images. A machine learning algorithm might be 97 percent accurate and a human might be 95 percent accurate, and you might think, OK have the machine do it. Actually, that would be wrong. You’re better off having the machine do it and then a human check it afterwards. Then you go from 97 percent to 99 percent accuracy because humans and machines make different kinds of mistakes.

But radiologists also consult with patients, coordinate care with other doctors, do all sorts of other things. Machine learning is pretty good at some of those tasks, like reading medical images; it’s not much help at all in comforting a patient or explaining the diagnosis to them.

MS: Which reminds me of the fiasco a while back where a hospital used a teleoperated robot to tell someone they were going to die and people freaked out. Well, duh . I don’t know why more roboticists aren’t warning about this. There’s certain jobs that humans will probably always do be good at, which are those that require the empathy that machines don’t have."

Read more here: