AI is having a moment right now.
Algorithms are everywhere: on our social feeds, in our playlists, on our highways, and in our inboxes.
Research shows that artificial intelligence is projected to add $15.7 Trillion to the GDP by 2030.
For businesses and consumers, it’s a huge win. We’re constantly looking for faster, cheaper, better and AI delivers.
For employees though, things are a little murkier. Automation makes work more efficient, but will it make things too efficient?
What will our jobs look like as AI grows smarter and more ubiquitous? Sure, the demand for data scientists and software engineers will increase, but what about the rest of the workforce?
Will robots and computer programs replace us?
Here are three reasons why there’s less to fear and more to celebrate.
1: Training Comes First
For all of the hype, AI is still very much in its infancy.
To get an AI program to the point where it can run independently and deliver accurate, impactful results, it needs first to be trained.
This requires a lot of human labor.
We’ve witnessed this first-hand at Ricoh Innovations with our retail execution solution. Our computer vision technology helps consumer packaged goods (CPG) brands and brick and mortar retailers efficiently track corrective actions (out of stocks, missing facings, incorrect placements) on shelves as well as understand longer-term trends like share-of-shelf and share-of-assortment to ultimately provide a better experience for shoppers.
For each product a user wants to track, our machine learning algorithm needs to be trained with images to recognize it. On average, this means the algorithm has to be fed at least 5–10 photos of the product in various angles, distances, and lighting conditions to replicate real-world conditions. And this number can be even higher when we’re looking at more complex products such as transparent, cylindrical, or deformable objects!
Multiply this against the average number of products any given CPG produces on a yearly basis, much less a seasonal basis, and the task starts to look pretty daunting.
This is why we’ve seen a steady rise of crowdsourcing training data companies such as Figure Eight and DefinedCrowd. These companies work with large teams of contractors to collect, label, and organize training data.
Relying on crowdsourcing allows AI companies to gather accurate real world, high-quality training data quickly and at scale.
As AI adoption continues to grow, we will see an increase in the need for people to fill these collection and annotation roles.
2: AI Is Not That Smart
AI is programmed to “think” according to a set of rules, which means that there are nuances that can be hard for the software to pick up.
Many times, AI isn’t capable of self-troubleshooting.
Take Tesla’s aspirations for a fully-automated manufacturing facility, for example. The Alien Dreadnought was projected to produce 5,000 Model 3 electric cars a week to keep up with growing demand. But instead, the factory kept falling short of about 3,000 vehicles a week.
The intelligent robots were creating bottlenecks rather than efficiencies.
In response, Elon Musk tweeted, “Yes, excessive automation at Tesla was a mistake. To be precise, my mistake. Humans are underrated.”
Musk then shared plans to add more human labor to the assembly line and grow to three shifts of manufacturing a day — this equates to hiring about 400 employees a week.
These new roles will require a lot of reskilling to get traditional manufacturing employees familiar with managing and operating AI-based robotic machinery.
As AI becomes more integrated into workflows, we will see more of this kind of job expansion in manufacturing and beyond.
3: Robots Can’t Keep It Real
Research shows that a majority of consumers (77%) chose, recommend, or pay more for a brand that provides a personalized experience.
But no matter how personalized chatbots and personal assistants become, nothing can replace that human touch.
We’ve seen some high profile chatbot fails over the past few years. We’ve seen bots respond with irrelevant or inappropriate answers, others which didn’t process information correctly, or even bots which communicated in an off-brand tone.
This is because, according to Ben Medlock of bigthink.com, AI is built very differently than the human brain. While the human brain is designed to predict, AI is a mechanical brain which is designed to memorize.
So while AI may replace some more tedious parts of our jobs, it isn’t even close to assuming the tasks that require experienced insight and judgment.
If anything, AI may be for work what ATMs were for banks back in the 1960s.
When ATMs were first introduced in the 1960s, bank tellers everywhere probably had the same reservations about this new automated technology putting them out of a job. But demand for human tellers grew. And 40-something years later, demand for bank tellers continues to grow faster than the labor force as a whole.
While ATMs reduced the number of tellers needed at each office, this lowered the average operating cost of a branch. This, in turn, grew the demand for branches, which drove up the demand for tellers.
In fact, we’ve seen this time and time again in history, where new technology seems to threaten specific occupations, but it ends up growing demand instead.
Now’s the time to embrace AI and familiarize yourself with how it might be integrated into your workflow in the future.
AI is coming, and it’s bringing more work, not less.