What career or professional decisions would you have made in 2018, if you’d known that in just two years, the world of work would look dramatically different?

There are no do-overs for us as prisoners of linear time, but we can use our pandemic-era experience of reinvention to prepare ourselves for the next major upheaval in the world of work.

And that upheaval may be just around the corner. In sharing a recent paper from a group of the world’s leading AI researchers, Managing AI Risks, co-author Geoffrey Hinton summed up the speed of this change: “Companies are planning to train models with 100x more computation than today’s state of the art, within 18 months.”

 

What AI acceleration means for your work

To get your mind around what that means, think about the difference between GPT-4 (the present model) and GPT-3.5 (the previous model, and the one that led to the world waking up to the suddenly awesome power of generative AI). GPT-4 is only 10 times more powerful than GPT-3.5according to OpenAI (the company behind GPT), and whenever I run out of GPT-4 credits and get bumped back down to 3.5, I feel like the AI has just been lobotomized.

So imagining an AI that is 100x more powerful than what we have today—well, it’s unimaginable. And yet, it’s right around the corner. It might take two or three or even five years before you see the impact of that increase in computing power in your own job or workplace, but it’s unlikely to take ten.

If it’s hard to get your mind around the world of work changing that quickly, remember that the pandemic changed the working world at a speed and scale we would not have believed, either.

So how do we prepare for the next unimaginable transition? Think about three potential futures, and then figure out a set of professional activities and investments that make sense for all three scenarios.

 

1. A future of continuity: Work mostly stays the same

Yes, AI is going to get more powerful and more pervasive. But maybe the world of work won’t change that much; maybe we’ll just adjust how we work a bit, and get a bit more done, and maybe some new job categories will emerge. Plenty of people point to the history of technological innovation and suggest that anxieties about mass displacement and unemployment will prove unwarranted.

If your work, employer and field remain substantially the same over the next decade, you can still benefit from making informed, incremental and effective use of AI tools. Look for opportunities to….

  • Reduce overload by using AI to triage incoming communications or digest big documents more quickly.
  • Automate routine tasks so that your work is less annoying.
  • Learn a new skill by using AI as a tutor, so that you can advance in your current career.

2. A future of upheaval: The world of work blows up

A massive increase in AI power could lead to a massive decrease in the number of available jobs—and this time, automation will displace not just blue-collar workers, but white-collar professionals like programmers and lawyers. As Daniel Susskind points out in his insightful and enormously readable book, A World Without Work:

Much of the current conversation about the future of work assumes that we only need worry when most people are left without work. But even in a world where just a sizeable minority of human beings—perhaps 15 or 20 percent—find themselves in that position, we should already be concerned about the instability that inaction might bring. Remember that in 1932, a rise in the German unemployment rate to 24 percent helped bring Hitler to power.

Yikes, right?

Setting aside the social ramifications for a moment, it’s worth thinking about what that kind of labor market upheaval would mean for your organization or career. Will you still have enough customers or clients to stay in business? Will your own skills remain relevant and remunerative?

Don’t try to predict the exact way your field or job might change, but also, don’t insist that your job or employer is safe just because it’s so damn terrifying to imagine it all getting turned upside-down. Instead, you might…

  • Use AI for some part of your work every single day, so you become adept at managing and working with AI.
  • Commit to trying one new AI tool in your field each month, so you can see how AI might reshape your own work.
  • Track evolving AI capacity by picking at least one AI newsletter or news source to read regularly. (I recommend Marshall Kirkpatrick’s AI Time to Impact, which is where I read about the research paper that inspired this newsletter.)
  • Identify a hobby or adjacent field where you could build skills by working with AI on a fun project or customized course, so you build an off-ramp in case that proves necessary.

3. A future of adjustment: We find new ways to work and share

We can imagine work changing dramatically in the next few years, without assuming that leads to social or economic apocalypse. As professionals, as managers and as citizens, we have the opportunity to shape the path of adjustment to AI so that automation leads to better workplaces, happier employees and perhaps even healthier communities.

Think about how we might manage the transition to a world where AI means that one person can do the work of a ten-person team. It’s not unrealistic to picture a single radiologist overseeing a team of AIs instead of a team of human doctors, or a single lawyer managing a team of AI contract-writers, or a single writer polishing up the TV scripts written by a “writers’ room” full of AIs. Maybe AIs will not do as good a job as humans, but then again, it’s not like humans do a perfect job in any of these fields, either.

Researchers are already forecasting where human headcount will be reduced or capped as more work gets assigned to AIs. At least some displacement is all but inevitable, because even if a legal team can produce 10x as much work with the help of AI, it may not be able to bring in 10x as much business. And if the law firm down the street is offering less expensive or quicker service by replacing some lawyers with AIs, it will be hard to resist the competitive pressure to do the same.

Automation doesn’t have to translate into mass layoffs, however, if we share both the pains and benefits of adjustment.

To imagine another path, think about the options a company might have if it can triple its productivity with the help of AI:

  1. Keep output constant, and lay off two-thirds of the workforce, increasing their profit margin by producing the same volume of products or services at a lower cost.
  2. Keep the workforce constant, and triple output and sales, increasing profits without increasing labor costs.
  3. Increase output and sales while keeping the workforce constant….but scale back each employees hours a little bit, without reducing their pay. The company increases profits, the employees maintain their income, and everyone gets a somewhat slower, more sustainable pace.

Sharing the gains from AI may sound out-of-step with mainstream businesses, particularly those holding onto the misplaced idea that they are obligated to maximize shareholder value. But there is plenty of historical precedent for employers taking a broader perspective on how to share the benefits of growth, from Henry Ford paying wages that ensured a stable workforce capable of buying the cars they manufactured, to the post-war compact that saw European unions, employers and governments negotiate a mutually beneficial approach to wages, working arrangements and productivity growth.

Now is the time to nudge our work, our organizations and our communities towards sharing the benefits as well as the costs of automation.  To put us on that path….

  • Propose win-win strategies where you increase your output with the help of AI, and see some of those time savings come back to you as reduced hours (but not pay).
  • Extend your role into new areas (where you’ve built skills with the help of AI) as the time required for other tasks shrinks.
  • Encourage AI experimentation by letting your direct reports know that they can share in AI productivity gains by getting more flexibility as they do their work in less time.
  • Talk with your colleagues about job sharing if overall head count is going down. Maybe some of you would prefer to work less, even if it means a pay cut.
  • Identify new services, products or revenue sources for your organization, built on extending output or offerings as AI changes what you can deliver.
  • Support unionization or collective advocacy efforts within your organization so you have a voice in how AI gets adopted and how roles get restructured.
  • Vote, volunteer and donate for candidates and policies that back compassionate adjustment strategies like guaranteed basic income, effective retraining programs and compensation for rights-holders whose work is incorporated into AI models.

Invest in all three futures

You could bet on just one of these scenarios, and focus your professional or political efforts on a single possible future. But it’s awfully hard to anticipate which path we are on, or how quickly we’ll accelerate.

That’s why it’s smarter to picture these three paths more like a Venn diagram. While there are some strategies that are only relevant in one scenario, you can invest in professional activities and social commitments that will be helpful no matter what the future holds.

So invest in ways that will be relevant if your work mostly stays the same, if all hell breaks loose or if you want to shape the way AI unfolds in your workplace, field or community. You can build AI skills that will make you more effective in the job you now have, develop knowledge or skills in adjacent (or wholly new) fields, and strengthen the relationships and capacity that can make our current organizations healthier and more inclusive, and lay the groundwork for a more compassionate transition if workplace displacement accelerates.

That other future

Admittedly, I have left out the worst-case scenario, which is the one that AI researchers warn about in the paper that kicked off this newsletter: “future AI systems could insert and then exploit security vulnerabilities to control the computer systems behind our communication, media, banking, supply-chains, militaries, and governments. In open conflict, AI systems could threaten with or use autonomous or biological weapons.”

As bad as that sounds, there is one up side to that scenario: You can leave it out of your career planning. After all, if the robots take over, you are definitely going to have bigger problems than what happens to your job.