This is NOT an article about how AI, automation, and robotics are coming for non-knowledge jobs. That is happening, but this is an article about how AI is coming for traditional knowledge economy jobs too, and how it will change our economy and society, and I think for the better!
A few days ago, I was watching NCIS with my mom (it’s always on some channel). As per usual, a clue came in, and with a few tips and taps on a computer they had traced it back to its source, cross-referenced it with a database, and sent the results to the field agents’ phones. In all, the scene lasted about 30 seconds. My mom said, “How can they do they so fast and by only typing? It takes me 20 minutes just to remember my password.”
NCIS is dramatized television. There are very few, if any, people or organizations with that level of computing sophistication and coding skill. However, it’s close enough to how we think computers work to be believable. More remarkably, we’ve at least thought we’ve been at the cusp of this level of computational sophistication for nearly 20 years. I remember watching 24 with my dad in the early 2000’s and very similar tip-tap-success was going on back then in that show. Yet we all know, by sheer fact of our daily lives, that working with digital information is cumbersome, time consuming, and does not always end in success.
Societally, we’ve convinced ourselves that we are living at the leading edge, if not the pinnacle, of the Digital Revolution. The advent of AI is just around the corner, and our 40+ years of digitization are poised to pay off into more leisure and more accurate and easy computing for all of us. On the contrary, I contend that we are merely at the beginning of the Digital Revolution, and there are still many years of work ahead of us before we can enjoy the tip-tap-success that we see on television.
Data remain very compartmentalized. Throughout the digital age, companies, governments, and other entities created databases, data protocols, and computing and data languages ad hoc. Even within large organizations different databases exist to house purportedly the same data, and sometimes these databases contradict each other. Furthermore, data are often user generated, so discrepancies propagate over time. Remember when they rolled out the electronic medical record (EMR) at your office and you could not find the field for pulse until someone told you to look for “heart rate”? And is the accounting system in dollars or thousand dollars? These are the discrepancies that real life NCIS confronts when they perform data analysis, and it takes far more time than we seem to think to sometimes get less than clear results.
A few weeks ago, I met my friend at a food hall in Midtown. I couldn’t help but look around and try to imagine what everyone did for work. Most patrons were young workers in business casual. Being Midtown, I imagined a lot of bank and finance workers, with a smattering of consultants, business people, and people in media and publishing. I know how they spend their days. I used to be a finance consultant. They spend all day pulling together data from disparate sources and collating them into something that their superiors can use to make decisions. “Where’s the data?” “Who has it?” “Is it any good?” These were my daily lines. Most work time for these “Midtowners” is spent replicating data, models, and results. Much less time is spent deciding what they make of results and numbers. I had entire projects where I figured out how the data came together and simply documented the process. Despite all of our advanced statistics and calculus classes, most people in these “Midtown” jobs are just performing basic arithmetic, if that.
However, there is reason to believe that the Digital Revolution will soon be accelerating. Emerging innovations like blockchain and internet of things (IoT) are streamlining the collection, storage, and sharing of data. The rate at which we generate data is accelerating, so having clear protocols for the sharing of data is key if we want to continue to move up the digital curve. If we continue to generate astonishing amounts of data but do nothing about their balkanization then making connections between data – the tip-tap-success we see on NCIS and 24 – will be more and more difficult, not easier, as we often assume as default for digital processes.
Over time, as fiefdoms of data come crashing down and the Digital Revolution truly does bring us closer to tip-tap-success, all of these Midtowners in clerical and finance roles will find themselves with a lot of free time on their hands (so will the consulting firms). Banks will finally be able to cut lose the throngs of high paid workers that spend their days knee deep in Excel, jockeying numbers for the few actual managers in firms whom make decisions. Managers will be able to easily retrieve <tip> the data they need, perform some manipulations as they see fit <tap>, and then make decisions based on their results <success>.
This Digital Revolution is a necessary prerequisite for the full advent of AI. Data are the fuel for AI. Machine learning algorithms require vast quantities of data, and preferably data that update in real time, so the algorithms can truly learn and improve upon themselves. As it stands now, all of the world’s data are too balkanized for machine learning algorithms to pull them in and turn them into the true putty that will lead to cognition-level algorithms. However, when it does, it’s not only the Midtowners that need to worry about their jobs. Managers – true actually make decisions of import managers, will begin to see their judgement challenged by algorithms. When there is little uncertainty in what has transpired in the past and what the forecast prognosticates for the future, there is little room for what we now think of as managerial judgement in decision making.
When I discuss this future with business people they see it as a hard pill to swallow. This is a natural response, but I’m apt to point out that there are excellent companies that are already working on AI for managerial decision making. As consumers, we are most familiar with Alexa or Google Home as voice-enabled personal digital assistants. However, Salesforce has Einstein, which helps sales and marketing teams with routine tasks. They’re already working on more advanced business applications for Einstein, and before long you’ll be able to ask Einstein, “Should we acquire a company or build a new capability in-house?” We are taught analytical frameworks to solve these questions in business school, so once we have the requisite data packaged into something that a machine can consume, why couldn’t, and why shouldn’t the machine answer the question for us (or even alert us to what questions we ought to be asking)?
[Salesforce CEO] Benioff even told analysts on a quarterly earnings call that he uses Einstein at weekly executive meetings to forecast results and settle arguments: “I will literally turn to Einstein in the meeting and say, ‘OK, Einstein, you’ve heard all of this, now what do you think?’ And Einstein will give me the over and under on the quarter and show me where we’re strong and where we’re weak, and sometimes it will point out a specific executive, which it has done in the last three quarters, and say that this executive is somebody who needs specific attention.”
I am not saying that the clerical workers of today are overpaid data jockeys not worth their weight in avocado toast. Nor am I criticizing their managers for hiring them and needing their assistance. I recall a particularly large project from my old consulting firm. It required an army of fresh-out-of-college consultants to comb through loan files and flag missing documents and other discrepancies. The work was tedious, but it required attention, occasional analytics, and downright intelligence. The young consultants did not find it particularly rewarding, only repetitive, and the bank certainly did not want to be paying the millions of dollars for error identification. Nevertheless, we still live at the dawn of the Digital Revolution, and this work was a necessary evil for everyone involved in the project. With AI far more popularized now than it was only ten years ago, nearly everyone can see the promise of AI in automating audit work like that. Nevertheless, it is still not a reality.
When the Digital Revolution does usher in true machine-powered cognition, I foresee banks, investment houses, insurance companies, and trading businesses, just to start, operating drastically differently from what we are used to today. Midtown will be cleared out – both the food halls and the corner suites. A few managers will rely on AI for most decision making, and the remaining workers will be more creative in nature, delving into new and emerging business models, or possibly still toiling in the age-old task of sales (with Einstein’s help, of course).
I hardly see this as apocalyptic for our knowledge economy. Yes, Midtown will be desolate, but Brooklyn will be bustling. Suit peddlers will be out of business, but hipster boutiques will be teeming. The advent of AI will be the advent of what I call the Creative Economy. Today, the creative economy is the corner of our economy focused on arts and leisure, design, media and entertainment, performing arts, fashion, and a smattering of other cottage industries.
Although people will lose jobs (or fewer new jobs will be created in the knowledge sector), our economy will be operating more efficiently. This relieves pressure on prices and leaves employed people with more disposable income. As an economy we can then deploy this disposable income into new interests, hobbies, passions, and arts. With more of the world’s most intelligent people free to devote themselves to their passions and leisure, there will be an explosion in creativity and creativity-as-commerce. Rather than focusing on high paying jobs devoid of meaning (if anyone who spends their whole day collating data says they “love their job,” they are lying), far more of our collective intellect can be dedicated to creative pursuits. We can create more content of an intellectual nature and consume more downright leisure.
I see four creative sectors staking claims for themselves and growing rapidly alongside AI:
- Pure Leisure & Arts
- Digital Arts
- Creative Enablement
- Physical-Digital Interaction
Pure Leisure & Arts
We already consider leisure & arts as very virtuous pursuits, although one that relegates all but the luckiest artists among us to being the perennial starving artist. These arts include writing, painting and drawing, film-making and acting, music, dance, fashion, gastronomy, architecture, other forms of literature, performing arts, and visual arts. With more time left to pursue the consumption of leisure or the practice of these arts, the traditional arts will proliferate.
Digital arts will be one of the fastest growing new creative pursuits. With more immersion in the form of augmented reality and virtual reality (AR/VR), there will be immense demand for graphic design, 3D design, animation, and VR environment design.
All of this art begs for software in which it can be designed, rendered, mixed, shared, and experienced. Today we have a knowledge economy, and Microsoft, along with the likes of SAP and Oracle, dominate knowledge software, so they are some of the largest companies in the world. In the future, the creative economy software manufacturers will be among the largest companies in the world. Design software by companies like Autodesk and Adobe will dominate our daily lives, and those companies will be vaulted into the Dow 30. There are even companies that are merging many technologies, from teleconferencing and virtual reality to design and architecture. They are creating software that will allow remote teams to interact in virtual reality environment and collaborate on design and creation real-time. Imagine a team of architects, all over the world, being able to virtually fly around the buildings they are designing and make changes together based on each other’s comments.
We will continue to live in a physical world (I am not predicting The Matrix), and manufacturing, engineering, medicine, and other physical sciences and fields will continue to be of the utmost importance. While less creative in nature, companies that bridge the physical-digital divide and allow AI and automation to assist in these fields will be extremely valuable. Importantly, they will continue to fuel the creative economy by freeing workers from tasks that can be performed by computer and machine, allowing them to more freely contribute to the creative economy.
The shift from a knowledge economy to a creative economy will have to be supported by the educational system. Training for trade and business will diminish. Instead, there will be more learning how to learn. Liberal arts will flourish, alongside an emphasis on mathematics and statistics, engineering, biology and medicine, and hard sciences. Coding, which is already moving to the mainstream of education, will gain even more importance, and humanities and the arts will once again be respected and valued fields of study. Education will also be prolongated and emphasized throughout one’s life, not just at its beginning, and there will be more economic emphasis on education. The creative economy will also self-reinforce the education sector by more effectively immersing learners in their education and create new and innovative ways to learn. If we can align our education system with the promises of the future and coordinate our data protocols for our collective well-being, the future will be bright, colorful, and fun and filled with enjoyable work and pleasurable leisurely pursuits.