Learn to Use AI at Work. Your Job May Depend on It. --- Office workers need to stay ahead of the machines that could replace them

"For the past two weeks, I've used cutting-edge artificial-intelligence tools in every aspect of my day-to-day existence, from my job to my personal life. Here's my verdict: The last time I had an experience this eye-opening and transformative was after I bought my first smartphone.

For most of us, these tools will enhance our productivity on tedious and time-consuming tasks. We'll be able to hand them off to generative AI, then easily check for errors. Regardless of your profession, the sooner you gain experience with using AI, the better off you will be. And it might just be vital to your employment.

In experimenting with AI, my aim was to get a handle on the impact it will have on the 100 million "knowledge workers" in the U.S. -- not to mention 900 million elsewhere in the world. That commitment included the research and writing of this column, which, for better or worse, would likely have taken a significantly different form without the help of AI. I didn't use AI to write any of the words you're reading now, but it did shape my thinking.

I plan to continue using AI to assist in my research, and in how I think through what I'm writing about. I would not be surprised if, someday, most knowledge workers find it just as indispensable.

After talking to some of the best (human) thinkers about the potential impact of AI on knowledge work, I'm convinced that we are now entering a new kind of industrial revolution, which many have begun calling the "Cognitive Revolution."

This nascent Cognitive Revolution -- the automation of knowledge work -- has important parallels to the early Industrial Revolution, when physical labor was automated. Those parallels could include higher overall productivity, and an increase in the world's total wealth.

But the Cognitive Revolution may also lead to millions of knowledge workers being pushed out of their jobs, especially those in roles that consist largely of tasks AI can do. By 2030, tasks that represent up to 30% of all hours currently worked in the U.S. could be automated by AI, according to a recent estimate by McKinsey.

Today's AIs can ingest documents, images and endless reams of instructions from their users, and then talk about them in a way that sounds like a person, but there's no mind behind the curtain animating that prose. It's all just a fuzzy approximation of the enormous amounts of data on which the AI was trained.

I've been using OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini and Nomi's and Perplexity's eponymous AIs.

 I've also come to rely on other programs' AI features, such as meeting transcription and summarization in Otter, and autocomplete in Google Docs, which has sped up my note-taking.

The best of these AIs -- the ones you have to pay to access -- are good advisers for tasks humans have done a million times before and written about ad nauseam on the internet. Asking GPT 4 for help with an ingredient substitution or advice on a simple weeknight recipe yields good results, and on numerous occasions saved me a lot of googling. 

It was equally capable of creating a marketing plan for a friend's small business.

The ability of today's large language models to customize their output is one of the main things that makes them better than conventional search, and passable at performing simple research or content-generation tasks that might otherwise be done by a person.

My experience shows how they could have a large impact on small businesses. Nathaniel Whittemore is the founder of Superintelligent, a platform for learning AI currently available by invitation only. He's found that the most popular courses his service offers are deeply practical ones, including ones that small businesses can use to multiply the productivity of whoever is in charge of their marketing. No one is expecting ad copy or Instagram posts to be works of art, and generating that kind of content, or even just rewriting or re-cutting existing content for different platforms, is a lot of work.

Small, resource-constrained companies and freelancers are precisely the folks who have been the earliest to embrace AI as a way to speed up basic knowledge work.

I spent about five minutes using a custom GPT in OpenAI's "GPT Store" (think of Apple's App Store, but for AIs) to generate a logo for an imaginary lifestyle brand. (If you read this and are inspired to create a clothing brand for middle-aged men called "Dad Life" with the tagline "Take My Pills / Pay My Bills," you owe me money.)

What does this mean for the graphic artist who would normally get hired to create my logo? I might not hire one because I can achieve a good-enough design on my own. But that doesn't mean graphic artists are all out of work. In fact, AI may streamline aspects of their jobs so they can focus on other things. 

One thing we can't automate away, for example, is taste, and we'll still need professionals who can steer AI, polish its output, and create reference materials to be fed into it in the first place.

Today's AI almost always automates individual tasks, not whole jobs. Some jobs consist mostly of tasks that can be automated, like customer service, content marketing and writing product listings for e-commerce services. There still has to be a person using and coordinating all those AIs, however, along with doing the abstract thinking that, for now, remains the sole domain of humans. This means that while AI isn't going to eliminate jobs, people using AI will -- which has been the pattern in automation since its earliest days.

I was able to use OpenAI's tool for creating custom GPTs to build my own custom assistant. My assistant is an example of the sort of "AI employees" that I think all knowledge workers will someday have at their disposal, to make their jobs easier. If you're already a ChatGPT subscriber, I invite you to try out "Tech Column Intern."" [1]

1. EXCHANGE --- Keywords: Learn to Use AI at Work. Your Job May Depend on It. --- Office workers need to stay ahead of the machines that could replace them. Mims, Christopher.  Wall Street Journal, Eastern edition; New York, N.Y.. 23 Mar 2024: B.2.