Top AI researchers share five insights on how AI will shape the future of work over the next five years

There’s no question that AI is reshaping the future of work. Nearly all executives—96%—now feel urgency to incorporate AI into business operations, according to the latest research from Slack’s Workforce Index survey of more than 10,000 global desk workers. And implementing AI is now a top executive focus, above concerns about inflation or the broader economy. 

The research shows that enthusiasm is warranted, as desk workers who are using AI tools aren’t just more productive—they’re also markedly more satisfied and engaged on the job. 

And yet, while AI adoption is increasing quarter over quarter, the Workforce Index also found that more than two-thirds of desk workers globally still have not tried AI tools. One major blocker is trust: Only 7% of desk workers consider the outputs of AI to be completely trustworthy for work-related tasks. 

So how should companies think about implementing AI, and what can leaders and teams do to build employee trust and confidence in these new tools? Slack’s Workforce Lab assembled a group of leading AI researchers for a candid conversation to tackle these questions and discuss how AI will affect the global workforce over the next five years.

Watch the full discussion here or read on for their top five insights.

AI will not replace people. People who are adept at using AI will replace people who are not.

My Stanford colleague Curtis Langlotz states, “Will AI replace radiologists?” is the wrong question. Instead, he states, “The right answer is radiologists who use AI will replace radiologists who do not.”

The big issue is we’re often not even solving the right problems—we do not know how to design AI systems to have a positive human impact. 

I believe there is a better way to design AI, and that better way is what I term “human-centered AI.” It starts with a user-centered design process that integrates well-known techniques from human-computer interaction and design that account for the needs and abilities of end users of computing systems. And then we need to rapidly iterate and improve a design through rigorous testing.

We need to creatively develop new designs that augment people—rather than replacing them—by accounting for their cognitive abilities and existing workflows. And we should base these on underlying theories about how and why people behave the way they do. This will help us develop AI systems that work for humans rather than futilely trying to replicate humans. – James Landay 

AI literacy is now essential for everyone.

AI literacy is core. AI is not going anywhere. Learning AI is like learning how to read. Just like when you teach your kids to read, you don’t stop once they learn how to read board books—they progress. This is how your AI literacy has to progress. We have to both understand what’s going on right now, because it’s rapidly changing, and we’re going to have to keep up with it. This is not a one and done situation.

Right now I see a lot of fear of AI which is not helpful. We can overcome the fear by realizing how much AI is already in our everyday lives. You’re using AI to find a picture of a friend on your phone, to estimate your commute time, or forecast what the weather is going to be. And the more those who are fearful can overcome the fear they have of AI and start integrating these tools and figuring out how they work, the more empowered they will be.  – Elizabeth Gerber

We need to tackle the problems that are currently unsolvable with “AI moonshots.”

My smart dishwasher has more processing power than the Apollo guidance computer of 1966, used for the first moon mission. And that enables me to start the rinse cycle from anywhere on the planet. And I hope you see that that is not a very powerful use of the technology, because even if I could start the cycle from anywhere on the planet, I’d still have to go into the laundry room to actually get the laundry out, right? 

So although this is really powerful, and I don’t have any problem with it, the thing is: my washing machine is never going to the moon. And so we use technology well not when we trivially improve or automate something we’re already doing, but when we take a moonshot—when we extend what we can currently do into things we currently cannot do.

So when we think about how to use AI well, we shouldn’t just be asking, “What human capabilities can we replicate?”  We actually have lots of humans with lots of capabilities.

We should be looking for moonshots. The best way to use this technology is to accomplish valuable work that currently can’t be accomplished, whether that’s addressing climate change, energy production, global food scarcity and the inefficiency of agriculture, increased access to education, the need for breakthroughs in medicine, and of course the need to do our work in more productive but also more creative, engaging and interesting ways. – David Autor

AI is a tool but you can also think of it as a new team member. 

Teams are the future of work—they have always been important, and they will always be important. The problems we’re solving today are so complex and dynamic, they require diverse expertise and perspectives. 

Less talked about is how you make great teams. People think just put a bunch of different people together and magic will happen. But that’s not true. Creating a strong team requires understanding what different parts of work need to be done, who’s good at what, who’s motivated to do what. 

In the future, the way we’re going to think about AI is: it’s just a new teammate. So just like we put great care into forming and molding a team, we need to think about AI as a teammate—what is it good at, and what is it not good at, how does it get along with other people, what do other people think of it, do they trust it? All those considerations that go into careful team creation, we need to think about with AI as well. – Elizabeth Gerber

Three simple exercises to build comfort with AI capabilities: 

1. Plan a vacation
2. 10 minutes a day of microlearning
3. Write your future bio

Plan a vacation

To understand AI’s potential and limitations in a low-stakes way, use a free AI tool to plan a vacation. Ask the AI to create an itinerary, specifying activities you enjoy. Through this exercise, you will encounter AI’s strengths, such as its ability to generate creative ideas, and limitations, such as its challenges with memory and understanding user preferences. – James Landay

10 minutes a day of microlearning

Our Workforce Index survey shows that many companies are not providing AI education or training to their employees. We also see from the data that employees who are trained in AI are 19 times more likely to find that AI helps them to be more productive. So we’re running an experiment at Slack where we’re providing 10 minutes of structured education a day for three weeks. It’s very lightweight, very low cost, and it helps people try their hand at ways to use AI—from fun ways to use it in your personal life, like writing a poem or a song, to work-specific ways, like using AI tools to compose copy or conduct research. – Christina Janzer

Write your future bio

Imagine what your professional bio will look like 10 years from now. Will you have built expertise in a specific field, traveled the world, written a book, or started a nonprofit? Use AI to generate this future bio for you. This exercise isn’t just about envisioning what’s possible; it’s about empowering you to see that you are in control of your future. By using AI to craft your future narrative, you shift the focus from technology dictating your path to you harnessing technology to shape the future of work, not just adapt to it. – Elizabeth Gerber