Next, Here’s How AI Is Transforming Work

Strategies for the age of AI

Key Takeaways:

• Generative AI (AI, genAI — let’s call it “toolAI”) is starting to impact certain work roles. Call center workers, freelance creatives, programmers: Your work is changing, and there may be less available for some kinds of work.
• toolAI is transforming some work. Work that can benefit from more effective task-chaining and problem-solving, like legal services and management consulting, is being transformed by reducing repetitive tasks and creating better workflows.
• toolAI is catalyzing new work. New roles like AI Engineers, AI Consultants: Your time has come.
• We need toolAI to augment humans, not replace them. This is the time where we need to double down on the uses of software to help humans continually adapt — and to ensure all humans have access to meaningful, well-paid work.

But remember, the three recurring impacts of automation are always:

• 1) The pace and scale of change. Many more are affected, and they are affected more rapidly than before. We need to become more adaptive humans.
• 2) Technology only automates tasks. Some jobs are lost, some new jobs are created, lots of work is transformed. We need work systems that empower and reward workers to adapt.
• 3) A global downturn will kill work before toolAI. The 2008 recession, and the self-inflicted Pandemic Recession, disrupted tens of millions of workers in the U.S. alone, far more than the projected impact of AI in the next ten years. But remember: Any major economic downturn always accelerates the automation of jobs, as employers “do more with less.”

--------------------

IF YOU WANT TO LEARN HOW TO LEARN with toolAI, I have two brand-new courses, which you can try for free. 1) Developing a Learning Mindset in the Age of AI, and 2) Learning with Agility in the Age of AI. But after you click, you need to at least start the course within 24 hours.

--------------------

AI IS TRANSFORMING (SOME) WORK

Good news first: Workers who leverage toolAI for activities like software development and content creation are seeing substantial leaps in their ability to knock out more work. In one study published in Harvard Business Review, “Gen AI reduced task completion time by an average of more than 40% across all participants.” And those who can use the tools are increasingly seeing an increase in pay: For example, the AI premium seems to be about +12% in pay for tech jobs. The tools can also be a boon for a variety of populations, such as being a game-changer for those who are differently-abled.

Also, a range of new work roles are being created, such as AI product manager, AI-powered cybersecurity expert, AI research scientist, and AI ethics officer. New in the C suite: Head of AI (and it’s even transforming some CEOs’ jobs). And, soon, you’ll see work roles advertised like AI Agent Designer.?

In fact, a study of millions of job listings suggests that demand for human skills is increasing in the age of AI.

Bad news next: Certain work roles are going to be deeply affected much more rapidly than others. Some call center workers, consultants, and programmers are already finding their work is being dramatically affected by toolAI. And LinkedIn chief economist Karin Kimbrough suggests that bank tellers, travel agents, data entry professionals, translators, writers, and marketers may also be impacted.?

For example, a recent report documents a significant negative effect on creative work, like freelance designers: At first there's a productivity bump, but then the loss of 30% of paid work, as employers use the same tools — and not the freelancers.

But some of the most vulnerable workers are the 19 million women in administrative, secretarial, and other support jobs. As their employers automate routine tasks, these workers may find their work shrinking — just as it did with the advent of the personal computer. (And, because of automation, there may be fewer career-path entry points for young workers and career changers.)

Other jobs that may be affected: Postal clerks. Payroll clerks. Movie dubbing. Translation work. And a new Bloomberg report suggests that Wall St. alone may lay off 200,00 workers. In fact, a lot of market predictors suggest that 10-15% of human work tasks could be performed by agents. (I’ll explain what an AI Agent is in another newsletter.)

But don’t be depressed. Nobel prize-winner Daron Acemoglu maintains that the truth will be much less dramatic, automating less than 5% of all work tasks in the next decade

THE DATA IS (NOT YET) TRUSTWORTHY — AND MAY NEVER BE

So why are these estimates all over the map? Because we’re all just guessing.

I’ve devoured every report I can find on toolAI+workforce research, such as OECD’s Job Creation and Local Economic Development 2024, Brookings’ Generative AI, the American worker, and the future of work, Workday’s AI IQ, LinkedIn’s AI At Work, JFF’s AI Ready Workforce, and many more.

But perhaps most influential is the World Economic Forum’s just-released Future of Jobs 2025. (WEF’s annual Davos meeting is this week.) Yes, WEF predicts 92 million jobs “could” be affected. But technological change is just one of the factors. Geopolitical tensions, green energy transition, economic shifts, and aging populations also have a big influence. And WEF also predicts 170 million new jobs to be created.

It’s not that one or more of these predictions may end up being true. It’s that a great failing of our work systems in the U.S. and beyond is that we don’t know. We clearly lack the tools to clearly understand the potential impacts of exponential technologies on work — and to provide workers and employers with toolsets to proactively adapt.

For example, an article by the contingent-worker company Adecco says that 13% of workers in 2024 “reported having lost their jobs” because of AI — without reporting a source. Dropbox announced it was laying off 500 people, “also citing AI” with no other detail. Outplacement specialists Challenger, Gray & Christmas said that in May 2024 3,900 people were displaced by AI at U.S. firms — again, without citing specific sources. That Wall St. layoff number? Also questionable.

Journalists, bloggers, and analysts too often contribute to the fuzziness of the information we receive. For example, grocery checkout kiosks and warehouse robots are automation, but they are not direct uses of toolAI software. Yes, relentless automation will relentlessly change a lot of jobs. But let’s not file it all under “AI”: That’s just literally making money for the tech companies that create the products

BEWARE SILICON VALLEY HYPE

And that’s what “AI” is: A product.

I have enormous respect for Kai-Fu Lee (KFL), former president of Google China. But when he predicted that AI would replace 50% of all jobs by 2027, the math doesn’t add up. Humans simply don’t adapt that quickly. His statement is an unfortunate example of the hype that benefits Silicon Valley companies and venture capitalists (which KFL has become) — designed to encourage a narrative of infallibility, so we’ll buy the products. And toolAI today is deeply fallible.

That’s why we need to beware headlines like this: 1X’s Eve humanoid robot masters task chaining, nears autonomous work. The article simply refers to a robot that may not need quite so much human oversight as it did before — but only if the engineers can keep it from making mistakes. (When was the last time your toolAI software didn’t “hallucinate”?)

Or this, from Axios: “Imagine a world where complex tasks aren't delegated to humans. Instead, they're executed with the precision, speed, and creativity you'd expect from a Ph.D.-level professional. We're talking about super-agents — AI tools designed to tackle messy, multilayered, real-world problems that human minds struggle to organize and conquer.”?

This year? Really? I’m sorry, it’s irresponsible reporting to use Silicon Valley’s breathless hype to suggest that these technologies will have any kind of general-purpose reliability any time soon.?

Look. “AI” is a marketing label. “AI” is a basket of technologies, one of which is generative AI, and a version of that uses LLMs (Large Language Models, like ChatGPT and its competitors). It's a deliberately-imprecise label designed to sound like it's everything and everywhere. There is no such thing as “an AI” — even if that's what you call it when the software shows up in your Zoom call.

And “an AI” never actually “takes” a job. If technology conducts most of the tasks that were performed in a job role, and the former worker gets laid off (especially when some software algorithm says they’re “a low performer”) — that’s pretty straightforward job displacement. But it’s usually not that simple. And a human is always the one making the decision that another human loses their job.??

“Nobody’s whacking 30% of their staff this year because they put in AI. Let’s just call it — that’s not happening,” says Maribel Lopez of Lopez Research. “And the reason that’s not happening is, let’s just say they wanted to whack 30% of their staff and put in AI. They probably still wouldn’t think it’s accurate enough.”

Well, except for many tech employers.

America already has the most productive workforce and economy the world has ever seen. Let that sink in. We already work really really hard, and we produce an incredible amount of work. Yet those who lead tech companies act as if that’s not enough.?

For example, the CEO of the startup Replit has made it clear he doesn’t care about coders. The poster child for AI job displacement, the buy-now-pay-later Swiss fintech Klarna, reported it will stop hiring humans (which it really isn’t — look at their online listings), and the CEO is “gloomy” because he thinks AI could take his job. But the company is preparing to go public, so its PR machine is on overdrive. (Wouldn’t a future shareholder love a company that sounds so automated?)?

Salesforce CEO Marc Benioff says they may not hire any more human engineers. And Meta CEO Mark Zuckerberg says he foresees a time when his company’s code will be created solely by software-based “AI Engineers” — a label that only a tech CEO could love.

As columnist Chris Westfall writes in Forbes, “AI is simply the enabler, a tool that allows companies to capture what banks are built on: profit.” And, increasingly, sadly, profit is what tech companies are built on, as well.

So why take it out on the workers? Amanda Marcotte in Salon has the answer: “...as with Elon Musk, Peter Thiel, Marc Andreessen and other tech billionaires, [it] isn't just about overgrown nerds trying (and failing) to compensate for high school insecurities. It also seems to be fueled by a deep loathing of the very people who have made these capitalists their money: the fleets of mostly desk workers who make the companies run.”

Couldn’t have said it better myself.

SLOWER IN THE SHORT TERM, GREATER IN THE LONG TERM

Even for technology as disruptive as toolAI and related products, history says that adoption is inevitably slower in the near term, but often exceeds predictions in the long term.?

It turns out that lots of work is remarkably resilient to automation, even for jobs you’d think would go away quickly. Many jobs are packed with a range of problems to solve, tasks to solve those problems, and human skills to perform those tasks. A technology may automate a lot of repetitive tasks, but the more complex and human-centric tasks often resist automation — such as the increasing demand for radiologists, a job that many predicted would be rapidly automated out of existence.?

As behavioral scientist Gregory Rosalsky quotes Stanford’s Erik Brynjolfsson in Planet Money , "According to my research, [radiologists] do about 27 distinct different tasks. One of them is reading medical images, but they also would do a lot of other things, including consulting doctors and patients. These are things that you wouldn't want a machine to be doing to people, and that's the way it is in almost every job in the economy."

Add human unwillingness to change, poorly-designed technology, and the resistance of labor unions, and a lot of technology takes longer to be adopted than we think. And we’re often wrong when we hear a few examples, and then project that impact on everyone in the same job: Just because there’s one “hologram Jesus” taking confession in a Swiss church, it doesn’t mean every priest will be out of work tomorrow.

And yes, you will see an increasing number of headlines about jobs lost to AI. Long-term, the impact may be as seismic as the Internet — or even greater. But in the near term, what you likely won’t see are many headlines about the new work being created, or about how existing workers’ work is being transformed.

But remember that work is mostly not about AI. In Western economies, about 60% of job roles like the trades have a strong physical component, which uses skills that are much harder to automate, so toolAI has little impact.?

THE PAST & FUTURE OF (YOUR) WORK: TECHNOLOGIES THAT EMPOWER HUMANS

As Galinda says in Wicked, "Why don't you just teach us history, instead of always harping on the past?"

To harp a little on the past, it can be helpful to look back 30 years ago at the impact of this new-fangled technology called the Internet. In 1995, I was the editor-in-chief of Inter@ctive Week, “the Internet’s first newspaper.” We said (frequently) that the Internet would change everything. Well, the McKinsey Global Institute’s just-released analysis of the “digital economy” (a.k.a. the Internet) estimates its impact at $8 trillion worldwide. That’s a lot. But it’s less than 8% of the entire global economy. So maybe it didn’t change “everything.”

That report also says that for every job “destroyed” by the Internet, 2.6 new jobs were created. Web site designer? Social media influencer? Three decades ago, nobody had ever heard of those work roles. You couldn’t email your grandmother. Yes, a huge amount has changed — but it’s taken 30 years for many of those changes. Sure, toolAI is in some ways moving faster. But we're still in the early days of the work transformation.

And remember, as we covered in last week’s newsletter, there are still many workers needed in lots of industries. In fact, given declining birth rates, many Western countries may soon reach “peak worker.”? What we need are not technologies to replace humans, but tools that augment them so they can solve new problems. Especially for people with less training, and with less access to meaningful, well-paid work. There is recent good research that reinforces what I’ve been suggesting for some time: toolAI and related technologies can help lower-skilled workers perform (some) higher-level tasks more effectively, and it’s to be hoped they will then make more money.?

But less-trained workers, those often doing the same thing over and over again, are still the most easily automated. “[AI is] going to increase the gap between the rich and the people who lose their jobs," Nobel prize winner Geoffrey Hinton suggests. "It's crazy, because we're doing something that should help everybody…but if the profits just go to the rich, that's going to make society worse."

In KFL’s defense, he has said the same thing: “A.I. is presenting us with an opportunity to rethink economic inequality on a global scale.” And as another Nobel prize-winner, Daron Acemoglu, says, “We need investment for alternative approaches to AI, and alternative technologies, those that I would say are more centered on making workers more productive, and providing better information to workers.”?

So what can you do Next?

• Stop believing the headlines. Let’s not freak out about all jobs evaporating. (Besides, the lack of critical thinking will kill us before AI. :)
• Pro-actively learn some toolAI products. Use them to solve real-world problems in your work, your learning, and in your life. You can even use them to intentionally design your next work role.
• For a macro view, look at sources like The Coming Wave by DeepMind co-founder Mustafa Suleyman , The Economics of Artificial Intelligence by Isaac Tham , and the National Academies’ Artificial Intelligence and the Future of Work, a good review of core issues for work and education from a science and policy lens.
• Use the power of the purse. Don’t buy products or services from companies that position their technologies as replacing humans. (And we haven’t even talked about the environmental impact of toolAI.)
• Don’t ask toolAI to harp on the past and teach you history. A new study said that even the best tool tested only got answers right 46% of the time — about the same as guessing, and worth a D- even if you’re grading on a curve.

What do we all need to do Next?

• Stop using Silicon Valley’s labels, such as “an AI.” Think of it this way: Every time you use that phrase, a bell rings, and you’re selling one more product for a tech company.
• Support U.S. legislation for better data, and more resources for those who will be displaced. The bipartisan house task force report on AI suggests the need for good data, but doesn’t mandate it. What could help solve that would be a national nonpartisan institute, spanning Labor, Education, Science, and other functions that could gather data from a range of sources, so we can understand the impact of exponential technologies on work and learning, beginning with AI.

Readers keep asking about AI, so in future newsletters, I’ll cover:

• Is AI really worth all the hype?
• The original sin of AI is human bias
• Raising the Robots: The machine in the cubicle next door — and in your parent's home
• What is human-centric HR in the age of AI?
• What’s AGI, artificial general intelligence — and does it matter?

And, what I’m absolutely the most passionate about:

• AI is not your co-worker, your employee, your workforce, or your co-founder

-gB Gary Bolles

I’m the author of The Next Rules of Work: The mindset, skillset, and toolset to lead your organization through uncertainty. I've served as the adjunct Global Fellow for Transformation for Singularity University. I have over 1.6 million learners for my courses on LinkedIn Learning. I'm a partner in the consulting firm Charrette LLC. I'm a senior advisor to aca.so, powering Skill-Building Networks for enterprises. I helped to catalyze Next CoLabs, a global thinktank of AI creatives, and I'm a co-founder of eParachute.com. I'm an original founder of SoCap Global, and the former editorial director of 6 tech magazines. Learn more at gbolles.com