Can AI usher in the 4-day workweek?

Are AI-generated productivity gains enough to mainstream a 4-day workweek?

Can AI usher in the 4-day workweek?
Photo by Lorenzo Herrera / Unsplash

Note: Smart Workweek and the Work Time Reduction Center of Excellence are hosting an event on August 16th with a panel of experts to explore AI and the 4-day workweek. Register here.


Bloomberg ran an article a few weeks ago with a headline: “ChatGPT Opens Door to Four-Day Week, Says Nobel Prize Winner.”

As someone who has rolled out a 4-day workweek at my company, and coached dozens of other leaders on how to make the switch, I’m skeptical that it will be artificial intelligence that will be this final constraint—that when unlocked—will usher in a 4-day workweek for wide swaths of the economy.

The Nobel Prize winner making this bold prediction is economist Christopher Pissarides of the London School of Economics. He says, “I’m very optimistic that we could increase productivity. We could increase our well-being generally from work and we could take off more leisure. We could move to a four-day week easily.”

In my experience, believing that you can transition to a 4-day workweek on productivity gains alone is like thinking you can become a millionaire by trimming your costs by 10%.

Productivity gains (through AI or otherwise) are necessary, but not sufficient to transition to a 4-day workweek. To make the transition, leaders need to work smarter, prioritize better, proactively structure their calendars, and continually practice disciplined focus on the most important things. Fortunately, AI tools can help.

The AI inflection point

There is new research predicting how AI will disrupt the economy:

  • A Goldman Sachs report found that AI automation could impact 300 million jobs globally and boost global GDP by 7% or $7 trillion per year.
  • A McKinsey report estimated that “current generative AI and other technologies have the potential to automate work activities that absorb 60 to 70 percent of employees’ time today.”

These predictions are well-founded. AI chatbots and large language models possess huge potential to generate major efficiency and productivity gains in the workplace.

  • Writing: For writing-heavy professions and roles, AI will increase the efficiency in the writing process (and help generate ideas for writers who are stuck). AI algorithms can now build upon ideas, edit text to make it sound more intelligent, parrot the writing style of famous writers, and distill complex ideas into simple prose. I’ve been playing with tools like to generate paragraphs and develop ideas.
  • Taking notes in meetings: I’ve begun using AI tools to take notes during Zoom meetings. When used well, AI meeting notetakers can lead to fewer meeting attendees (you no longer need a note-taker), and people who didn’t attend can catch up on video highlights, key takeaways, and next steps.
  • Accelerated coding: AI models can accelerate the speed of writing code. ARK Invest predicted that AI could lead to a 10x increase in coding productivity as developers partner with large language models to write and fix code at record speeds.
  • Customer support: AI will likely usher in major changes to customer support roles. Some roles might be eliminated, but AI will also make customer support even better for customers and companies alike. The Harvard Business Review released a report earlier this year analyzing how AI will change customer support and how it will put a greater premium on human expertise.
  • AI Agents & automation: People can give AI agents, like AutoGPT, an objective, and it can break down the tasks necessary to complete it. By connecting AI agents together, complex tasks can be accomplished: from fully building a website to fully creating a startup with $100.

Teams that are able to leverage best practices in prompting AI will be able to use AI tools to speed up dozens of processes and tasks within organizations. The modern workplace is already full of opportunities for human-computer partnerships to complete tasks, and new AI tools will take this to the next level.

Increasing productivity isn’t a silver bullet

But will all these advances in AI be enough to increase workplace productivity by 20% (from five days to four)?

Historical data say it’s unlikely.

Between 2010 and 2019—a decade that saw major technological advancements—productivity in the United States grew more slowly than in any other decade of the post-World War II era, at only 1.1% per year.

With all of today’s technologies that promise efficiency, automation, productivity, and collaboration, the average American worker spends 250% more hours in meetings each week than before the remote work transition.

For teams that are already working hard on short deadlines on a five-day workweek, it’s almost an insult for a leader to tell their team that they want to see each person increase their productivity by 20%, even if each employee is armed with the latest chatbots.

Employees might respond by saying that if the company is really serious about a 4-day workweek, then maybe greater strategic clarity, fewer meetings, less reactive firefighting, and better boundaries with clients would all be better places to start.

Prioritization vs. Productivity

If I had to choose between making my company more productive or making my company better at prioritizing the most important work, I would choose better prioritization every time.

In my experience, many organizations are overloaded with too many meetings, too many priorities, and too much work that doesn’t directly drive the company forward.

Moving to a 4-day workweek was hard for me as the CEO of Uncharted because I was a brute-force entrepreneur. For years, I focused on doing whatever it took to get the job done, which meant I hadn’t developed the muscle of careful prioritizing. But shifting to a 4-day workweek was a powerful forcing function for our team to get smart about distinguishing what was truly essential from what wasn’t.

We refocused our strategic plan on fewer priorities, reduced the number of meetings, and revisited expectations and norms we set with our partners.

Instead of getting faster, we tried to get smarter. Our biggest gains came from subtracting non-essential work, not from boosting our productivity.

How to leverage ChatGPT for the 4-day workweek

The best way for leaders to leverage AI to move to a 4-day workweek is by using it not to get faster, but to get smarter. I’ve identified two applications for ChatGPT that can help teams get smarter (there will be many more to come).

Application #1: Use AI to get smarter about saying yes and no

It's far easier to say yes than it is to say no. When someone asks us to review a document, when we get an unsolicited email, when we are asked to join a new committee at work, it's often tempting to say yes without thinking too hard. But when we say yes to too many things…

  1. We over-stretch ourselves by over-committing, and we burn ourselves out by doing it all
  2. We begin to drop balls and run the risk of not doing any one thing well because we've taken on too much
  3. We don't refine the skill of saying no, which leads our prioritization muscles to atrophy

The more we say yes, the harder it becomes to say no. We get out of practice.

There's a lot of fear in saying no. "What happens if people won't like us?" "What happens if by saying no, I will be looked over for the promotion?" "My reputation is all I have, and if I say no, then my reputation will be threatened." So much of our fear around saying no is connected to our reputation and being liked.

So if we can find a way to say no that keeps our reputation intact, then it will become easier to say no and we can stay focused on the most important work before us.

I’ve found that often the biggest barrier to saying no is the lack of the right language. If we had the confidence that we could deliver a no gracefully and it would be well received, we might do it more often.

Consider customizing the following ChatGPT prompt to get better at saying no. Here’s a prompt I have been using recently:

Instruction: Generate three options of a [length] response to [email] that respectfully declines the request/opportunity outlined in the [email]. In the response, take into account the [background context] and use the [tone] so that the [outcome] is most likely to be created.
Length: [define the length of your response]
Tone: [Define the tone of your response]
Background context: [provide the background context so ChatGPT knows what context to work within]
Outcome: [generate the outcome you wish to effect by sending your response]
Email: [copy email you received]

Recommendation #2: Use AI to get smarter about decision-making

One of the most important skills a leader can develop is to improve their ability to make good decisions. AI can help by being an immediate strategic thought-partner by pointing out blindspots, sharing different perspectives, and anticipating unexpected outcomes. For leaders looking to transition to a 4-day workweek, it can help identify what not to work on.

I’ve created a ChatGPT template below that can be used for each of the following exercises around getting smarter with decision-making:

  • You have a proposed plan of action, but you want to anticipate what push-back, concerns, or questions would be from your team or for a client with a given plan or decision.
  • You are considering making a decision, but you want to generate an analysis of that decision from a different perspective (for example, you might want thought-partnership on the impacts of this decision 6 months from now by taking a long-term perspective).
  • You are considering making a decision, and you know that you tend to have certain blind spots or biases, so you want an analysis of the decision that doesn’t have those biases or blind spots.
  • You are looking to identify patterns or insights across a set of historical results (measured in data). You have a sense of what patterns exist, but you’re looking for a different analysis to identify anything you’ve missed.
  • You have a roll-out plan for a new project, and you’re trying to anticipate all the ways the project could go wrong, so you’re looking for a list of anticipated challenges or issues that could impede its success.

Use the following template and build out the fields based on the specific strategic thought-partnership and analysis you’re looking for.

Instruction: Review the [proposal/strategy/plan] and generate a [length] [output] with the following [output characteristics], while incorporating the [background context].
Proposal/strategy/plan: [insert]
Length: [insert]
Output: [insert]
Output characteristics: [insert]
Background context: [insert]

Here are a few examples:

Anticipating pushback and concerns


Tip of the iceberg

The computational power of artificial intelligence is doubling every six months, which means there will be dozens of new applications for how AI can increase productivity, make teams smarter, and support the transition to the 4-day workweek.

But if history is any indication, the newest technologies often don’t have an immediate impact. The automated telephone switching system, which was invented in 1892, replaced human operators. But even into the mid 20th century, there were 350,000 telephone operators in the US. It wasn’t until nine decades after the invention of the automated switching system that the occupation of telephone operator largely disappeared in the 1980s.

The hype around AI right now makes it seem like its impact will be immediately felt, but it’s possible that it will take time for teams—and industries—to use it to its full advantage.

I’d love to hear from you about ways you’re using AI to work less.