AI Talk: Gig worker, Sidewalk and SpaceX

December 18th, 2020 / By V. “Juggy” Jagannathan, PhD

This week’s AI Talk…

Gig Worker

I saw this article in MIT Technology Review this week about the plight of the gig worker. The article features an interview with Professor Saiph Savage of West Virginia University (where I happen to be an adjunct faculty). Professor Savage has impressive credentials! She was named one of MIT Tech Review’s 35 Innovators Under 35. She works at the intersection of AI, Human Computer Interaction (HCI) and ethics. In this interview, she exposes the underbelly of the AI revolution: The hordes of people working below minimum wage for platforms such as Mechanical Turk.

What exactly is this platform? It allows tech companies to post work only a human can do which are of short duration and can be done usually with some basic proficiency. Everything is done online: qualifying to bid for the job, the job itself and being paid for it as well. So, what type of work have these tech companies been hiring out? Machine learning requires lots and lots of labeled training data. If you are building a sentiment detection program, you need to have a lot of sentences that can be labeled as having a positive sentiment or negative sentiment, for instance.

In the article, Professor Savage indicates that she believes having this type of work available is a good thing. I was surprised to learn that these platforms (there are many) are estimated to provide work for a million people or more. But there is a serious problem. The real wages as measured by hours spent and amount earned shows convincingly that very few earn even minimum wage, which clearly is not right. Professor Savage’s team has developed tools to help such workers determine which of these jobs to take and which to refuse in an effort to improve their lot. Tech companies who hire out these jobs should bear the brunt of responsibility for fair wages and need to make a concerted effort to rectify this wrong. More power to civic minded professionals such as Professor Savage!

Sidewalk from Amazon

I came across this article in Geekwire about Amazon’s new tech solution Sidewalk. What problem does it solve? Turns out, our “always on” culture, with a multitude of devices all connected to the internet, needs 24/7 connectivity. How can one ensure connectivity to all the devices all the time? Why not steal some bandwidth from other devices which have connectivity nearby? The aptly named solution, Sidewalk, is a metaphor for the sidewalk in our yards provided as a convenience for community use. This “sidewalk” is the same. Any internet of things (IoT) device, in this case from Amazon, will be able to utilize bandwidth from any other Amazon device (notice there is no mention of who owns these devices).

So how does this work? Forced opt-in is the mechanism at play. You are notified that a certain portion of the bandwidth of the devices you own will be utilized to serve the community’s needs. You can opt out, but it will require some serious digging to figure out how.

It turns out Amazon is not the first to try this strategy. Comcast Xfinity has a wildly successful implementation of the same strategy. Xfinitiwifi, which I am sure everyone has seen on their smartphone at some time or other, is a free wifi network for Comcast customers “stealing” bandwith from other customers. They have also adopted forced opt-in and conveniently do not advertise this fact. I have been supporting this program probably ever since its inception – blissfully ignorant of it! At least Amazon is trying to be a bit more upfront about it, perhaps because they are under a lot more scrutiny than Comcast.

SpaceX Spectacular Test

Leave it to Elon Musk, the master showman and uber entrepreneur, to put together a spectacular, groundbreaking test. The test? How to get a Starship super-rocket to high altitude and bring it back in one piece. This rocket is a prototype of one that can be used to deploy satellites, point-to-point terrestrial travel, to moon landings, to trips to Mars!

In the test, the super-rocket behaved as envisioned—reaching the height of 7.8 miles in a vertical lift off. The ending was spectacular as well. Check out this five-minute video of the whole event. It returned to the designated point, righted itself and slowed the descent. However, that process did not work completely as designed. The rocket crashed to the ground and went up in a ball of flame, euphemistically called Rapid Unplanned Disassembly (RUD). Elon tweeted the core mission was successful and there was a congratulatory note from rival Jeff Bezos as well.

AI Talk will be back online in January 2021. Wishing readers a Happy Holidays and a Happy New Year!

I am always looking for feedback and if you would like me to cover a story, please let me know. “See something, say something!” Leave me a comment below or ask a question on my blogger profile page.

V. “Juggy” Jagannathan, PhD, is Director of Research for 3M M*Modal and is an AI Evangelist with four decades of experience in AI and Computer Science research.