MIT Professor: Will AI impact jobs?

The weekly newsletter of the Business of Tech, giving you new insights into the world of IT service delivery.

Looking for stories from the podcast stories? Check out the pod itself on Apple Podcasts, Spotify, or daily in your inbox. Stories are available to everyone for five days,and Patreon supporters forever.

Was this forwarded to you? Join the list!

Will AI take the jobs away?

As we continue our exploration of various takes on the past, present, and future of AI, I keep coming back to a conversation I had with Dr. Yossi Sheffi, Director of the MIT Center for Transportation and Logistics at MIT.

Our conversation brought together two perspectives that I found to be rather grounding: an academic, unbiased look at the big picture of technology and an industry-specific peek behind the curtain of how AI’s tangible applications throughout the supply chain.

In need of an AI reality check? Me too. Here’s what Dr. Sheffi shared on a bonus episode of the Business of Tech.

An Academic’s Historical Perspective

As a professor of Engineering Systems, Dr. Sheffi has a robust understanding of systems optimization and risk analysis – both in the present moment, and throughout history.

So when I asked him what he thinks of the buzzy headlines promising that AI will either destroy jobs or accelerate businesses, it’s no surprise that he turned to the industrialization revolution as a case study. He highlighted that during this comparable period of technological upheaval, surprisingly few jobs were actually lost, many more were created, and a surprisingly large number of roles stayed exactly the same.

And even for the jobs that were destroyed, it took a lot longer than you might think for them to actually go away. He used the example of telephone operators: although the automatic telephone exchange was invented in the 1890s, there were still 350,000 operator jobs in the United States in the 1950s.

So no, Dr. Sheffi doesn’t think that jobs will go away any time soon, if at all. Here are some reasons why:

Generative AI is not perfect. It makes mistakes, and those mistakes are too risky for most businesses. It’s trained on human data after all, and humans aren’t perfect either.
Labor unions just won’t allow it to happen.
Governments aren’t jumping up and down at AI adoption, and likely won’t for quite some time. This means regulation won’t trickle down for a while, either.
General acceptance will hold adoption back. He used two solid examples here: when elevator operators striked in 1945, New York City ground to a halt because folks were scared to go on them alone. And, though modern planes can go gateway to gateway without a pilot, no one would get on a pilot-less plane.
AI has already been used for years. Robots in warehouses and Google Maps predictions are things we use every day, and there wasn’t a massive labor panic about those.

In short, when people approach Dr. Sheffi on the brink of panic, he says this:

“Well, first of all, come down, take a cup of tea, then we’ll talk.”

Shifting Focus to Tool Deployment

If you’ve been tracking my own takes on AI, you already know that Dr. Sheffi’s POV validates a lot of my own thinking. Again, I’m in the same school of thought as the folks at Microsoft pushing the Copilot language. To me, it’s about enabling effective work, not full replacement.

Dr. Sheffi provided another great historical example along these lines: Spreadsheet. Before its release, programmers spent months producing the same work you can now generate in seconds. When it showed up on the academic scene, professors like himself quickly switched from teaching students how to do the manual work to teaching them how to use it, use it well, and monitor the results.

In plain terms, he agrees that it’s just a tool.

Now, how can we start helping customers deploy this tool effectively?

Dr. Sheffi has such a great perspective on this query because of his expertise in warehouse automation, which is an area where AI adoption has been firing off for years now. Again, he pointed to a solid example from the not-so-distant past – even though Amazon has been a leader in warehouse automation with the Kiva robot, the company has since hired 1.2 million more people. Because they got more business, they were able to do more, and needed more people to do it.

While watching his own industry adapt to this technology, he noted this:

“People have one huge advantage to understand context, which AI does not, because they just look at the next word and the next word. There’s no context.… It’s a very systematic way of doing probability and working off a list of words… being able to judge [the results] is a skill that will be in very high demand in the future.”

Training Customers to Leverage AI Effectively

People in Dr. Sheffi and my school of thought are more of the belief that the biggest impact of AI will be how we learn to use it (not that it’s going to replace us). When I asked him what we should be teaching customers and teammates to do so effectively, he pointed out that simply monitoring an AI’s work won’t cut it.

Why? Again, he provided an example, this time from the present: self-driving cars. In theory, you can tell a driver to keep an eye on the car’s performance, but in practice, it’s far too easy to doze off and miss something important. The same thing will happen if you ask your people to simply watch AI all day.

So, Dr. Sheffi’s advice on this front is two-fold.

First, you need to keep people up to date on the manual way of doing things. His example for this one was Russia’s 2017 cyberattack on Ukraine, when Maersk’s computers went dark. Things got back up thanks to the people on the ground who could solve things by hand and fax the solutions around.

This is something we, as tech service providers, can start thinking about – training folks on the fundamentals of a task alongside the AI method. He believes that now is the time to start looking at training software, starting with smaller tasks.

The second piece of advice he offered is a bit more big picture; he wants governments, universities, and companies to plan ahead for new methods of educating younger workers who might enter the workforce into a job that’s newly automated. He could see a new style of the apprenticeship coming into play, where young folks work in AI-impacted roles after high school with a hybrid college program on the side.

A fascinating premise, to say the least.

Feeling a bit more tethered to planet Earth? If you enjoyed Dr. Sheffi’s on-the-ground and academic perspectives, check out his new book The Magic Conveyor Belt – Supply Chain AI and the Future of Work.

The AI conversation is far from over, however, so be sure to send any lingering thoughts and AI musings my way: [email protected].

More from MSP Radio

Missed Things?

How about our latest videos to catch you up?

The Daily Podcast available as videos

Marketplaces, Influencers, Pax8 and SMBs with Shirish Nadkarni

Strategy in the Digital Age: Digital Transformation made practical with Michael Lenox

Limiting Hallucinations, Category Theory, and predicting policy with Eric Daimler

How AI is different and how to link value in digital transformation with Kamales Lardi

Want the Daily News?

All the stories from the daily Business of Tech Podcast are available in the daily digest, and stories are available to everyone for the first five days, and Patreon supporters forever. Catch the audio of the show anytime on Apple Podcasts, Spotify, YouTube, or wherever you find podcasts. Links at businessof.tech

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.