This site is in beta. Tell us what you think.
Chapter 1 | Machine Coworkers Guidebook

Introduction to Machine Coworkers

Rather than focus on the potential for automation of rote tasks, organizations are increasingly looking for ways that machines can be used to augment human intelligence. When a machine is used to augment our own capabilities rather than replace them, we can think of it as a 'machine coworker'. The idea of working with machines as collaborators rather than straightforward tools raises some interesting questions.

How do we think about, coordinate, and partner with machine coworkers? What kind of work is best suited for the union of machines and humans? And what could we create if we had more 'human' capacity in our businesses?

Shifting from automation to augmentation

Automation is only one way of thinking about how machines can contribute to productivity and work. It is shaped by its origin in the industrial era and the idea of mechanized assembly lines and basic robots, in a time before machines could 'think' and 'learn.' Automation is the replication and repetition of human work processes using machines. For example, an 'autofill' tool in spreadsheet software can automate the task of creating a sequence of numbers (such as 10, 20, 30) so a human doesn't have to manually enter them.

There are two main limitations to the automation mindset:

  • It's actually quite hard to automate some parts of human work, especially communication work, emotional intelligence, pattern recognition, and making decisions in 'gray areas' or unknown fields.
  • People don't like to think about automation because they are worried machines will replace them which would lead to losing their jobs (and thus livelihood).

If we're able to shift our thinking from automation to augmentation, our concerns about obsolescence become less pressing and we can start to see real value for us as workers in having machines that increase our own capabilities.

Augmentation is a more realistic term for the current and near-future capabilities of machines vis a vis 'professional' or 'skilled' work. Machine coworkers can expand human capacity, improve quality, reduce stress, increase speed, and help humans focus on the most human tasks. A related shift in thinking is moving from Artificial Intelligence (AI) to Intelligence Augmentation (IA). For example, content recommendation tools can augment a researcher's fact-finding work so that they can spend more of their time synthesizing new ideas.

But what about human jobs?

If we're being honest with ourselves, we know that some human jobs will go away as technology continues to evolve. We can also assume that new jobs will arise due to that same evolution. In the long run, there is a possibility to drastically reduce the amount of functional and mental energy humans spend on work that machines could do. One concern that is important to acknowledge is that the future of work and machine coworkers will not arrive at the same time for everyone, and may not apply evenly or fairly. Very uncomfortable changes may occur for significant parts of the global population. At the same time, economists and technological theorists predict that ultimately, quality of life, life expectancy, productivity, and efficiency will increase for most if not all people through machine coworkers. Machines may even help us solve seemingly intractable problems like hunger and over-consumption of polluting energy sources. As with other digital transformations, new tools are not enough. We must also shift our thinking, learn how data works, understand new business models, and acquire new skills to navigate digitization's existential and organizational effects.

Photo: A robot's arms are poised over a keyboard as if it were about to play music with the text "The future of work and machine coworkers will not arrive at the same time for everyone, and it may not apply evenly or fairly."

The future of work and machine coworkers will not arrive at the same time for everyone, and it may not apply evenly or fairly.

To explore the potential of machine coworkers, we also have to acknowledge the fear that automation brings up. In a fear mindset, humans often focus on what they’re afraid of—even if they are bracing against it. In the context of automation and technology in general, many professionals bring up concerns about how they can stay relevant. Focusing on relevance keeps you working to convince others (be they customers or employers) that you or your company is capable to handle what's happening right now—but it keeps you from looking forward to what you can create next.

If possible, try to reframe concerns that come up for you around human jobs and machine coworkers as opportunities for unlearning—challenging your beliefs and assumptions about work and value. If we unlearn that there is always more work to be done than we have time available, and unlearn the mindset that most work can only be done by humans, interesting questions arise.

What will we do with the time we save on rote work? What distinctly human jobs can we pay more attention to now that those tasks are being handled by machines?

The key is to remember that when new technologies are introduced we need new mental models to go with them, otherwise it can be very difficult for people to consider how they want to engage with the changes those technologies will bring from an empowered perspective.

Here's one mental model that is helpful when addressing fear of the unknown or of obsolescence: the antidote to fear in the context of technology is a clear sense of purpose. When evaluating technology, roles, and decisions from the context of purpose, new opportunities become visible and we begin focusing on what we want the future to be like, rather than what we are afraid it will mean for us.

A young girl looks directly at the camera, her hand is resting in the hand of a robot. With a text box that reads "Many economists agree that coworkers will cause loss of human jobs... but also create many others."

Many economists agree that machine coworkers will cause loss of human jobs… but also create many others.


Problem types and machine coworkers

The type of problem will determine the feasibility and type of machine coworker that can help:

  • In a simple problem, we understand both what the problem is and how to solve it.
  • In a complex problem, we understand what the problem is but not how to solve it.
  • In a wicked problem, we understand neither what the problem is nor how to solve it.

Simple problems are often the easiest to find or create a machine coworker for. Simple problems also tend to be low-value work in human terms. For example, we know what data needs to be entered into a system and we have access to the data but it takes too long to enter it and we're always trying to catch up.  What we need is more 'people' or a faster process. A machine coworker can give us both.

However, just because a problem seems simple to humans doesn’t mean it’s simple from a machine's perspective. For example, driving a car seems simple to us once we’ve learned it because the mental models, knowledge, and skills required are relegated to our unconscious. If the problem cannot be turned into something a machine can solve through computational thinking and current machine 'learning' technologies, then the job must stay with a human. This is one of the reasons that thinking of machines as coworkers can be an epiphany—the 'machine talent' you want to 'hire' has to exist (or you have to train it).

Exploring machine coworkers by thinking styles

INFOGRAPHIC: a rectangular shaped field with eight squares that represent concepts which are arranged according to a X and Y axis . The Y axis is titled Orientation and the X axis is titled Focus. The concepts are explained in the body text of the article which follows this graphic. 

The eight squares are grouped in four columns with two rows. Two themes organize the rows on the Y axis from left to right. The topmost row is titled Big Picture and the four thinking styles from left to right represented are Explorer, Planner, Energizer, and Connector. 
The lower row from left to right is titled Details and has the thinking styles of Expert, Optimizer, Producer and Coach. The X axis has four columns from left to right Ideas, Process, Action and Relationships. 

Each thinking style is contained within the intersection of an orientation theme and a Focus theme on a grid. For example the intersection of Big Picture and Action is the thinking style of Energizer. Or, the thinking style at the intersection of Details and Relationships is Coach.

One way to think about different kinds of 'machine coworkers' is to consider what kind of thinking they need to perform. The Thinking Styles model suggests that there are eight fundamental kinds of thinking often applied in the workplace. As suggested in that article, it can be useful to consider which thinking styles are in your individual 'genius zone' versus others you are competent at (but unexcited by) or struggling with. Human colleagues can round out our thinking styles—and, in some ways, so can machine coworkers. For example, a program that looks for patterns in workflows and makes recommendations to increase efficiency could bring Optimizer thinking to your team. Or one that monitors news channels and collects breaking stories on topics of interest to you would be doing Explorer thinking on your behalf. When you notice some missing thinking styles in your team or organization, consider whether machine coworkers could fill in the gaps.

Explorer Thinking with Machines

Explorer thinking is about generating creative ideas

Machines can help with trend forecasting, ideation, pattern identification, and cross-referencing data points. And digital 'brains' can handle massive quantities of data.  In partnership with humans, the ever-increasing capabilities of machines to learn, in both supervised and unsupervised modes, makes it possible to sort this data into meaningful information. Machines, if properly instructed, can also watch and correct for some types of erroneous interpretation to help their human counterparts avoid bias (or at least identify under- or over-represented groups in data sets).


Planner Thinking with Machines

Planner thinking is about designing effective systems

Machine coworkers can help us design and manage complexity. For example, computer-aided design (CAD) is invaluable for complex engineering work, and project management software can help us stay on track without having to know all the moving parts of a workflow. More advanced machine coworkers that 'learn' and 'think' (machine learning and artificial intelligence) can even help us find patterns or narrow variables in very challenging systems, like the human genome or traffic patterns.


Energizer Thinking with Machines

Energizer thinking is about mobilizing people into action

It's hard to imagine machines as cheerleaders, capable of inspiring and mobilizing people into action. Yet teams are already starting to have machine 'bots' in their collaborative online workspaces that encourage good work habits. For example, the workplace chat app Slack has a plug-in that reminds human users to update each other with their work plan for the day, and sentiment analysis bots like the one in transcription software otter.ai can capture questions or action items that arose during a meeting without being prompted by a human attendee.


Connector Thinking with Machines

Connector thinking is about building and strengthening relationships

When you imagine what this thinking style is like, you might think of a recruiter or champion networker. Machine coworkers can help us connect too. For example, when a social network recommends new connections (or connections to rekindle) or smart customer relationship management software suggests news articles to share with your top contacts, those are machine coworkers who are good at connecting people to each other.


Expert Thinking with Machines

Expert thinking is about achieving objectivity and insight

Objectivity is something digital machines can be very good at. When you use a search engine, you're leveraging the expert thinking of its algorithms to sort through all the possible search results and rank them by what's most likely to be relevant. When you carefully analyze and search data in spreadsheets to find outliers and key data points, you have a research assistant helping you that just happens to be computer code. With machine learning and AI, machine coworkers have evolved from helping us organize our own research to suggesting new lines of inquiry and surfacing hidden correlations for us.


Optimizer Thinking with Machines

Optimizer thinking is about improving productivity and efficiency

If there's one thing machine coworkers 'love,' it's twiddling the knobs and adjusting the levers of our digital systems. Website analytics software, grammar-checkers, multivariate testing tools, and e-mail spam filters are all examples of functions that used to be substantial drains on human focus but which are now handled by machine coworkers.


Producer Thinking with Machines

Producer thinking is about achieving completion and momentum

In the modern workplace data entry, transcription, and many other business functions are commonly supported by machine coworkers well-suited to repetitive work. In fact, this is often the most obvious application of machine augmentation and automation. But even producer capabilities are evolving in exciting new ways such as machines that can synthesize natural language and even images, algorithmically creating news articles and artwork.


Coach Thinking with Machines

Coach thinking is about cultivating people and potential

Coaches are people who encourage you and give you valuable feedback and advice to help you succeed. Like those human partners, machines can also help us navigate decisions and cultivate our potential. Decision-support tools (like a 'wizard' for picking out a certain product or narrowing down a medical diagnosis) and recommendation engines for learning and personal development are just two examples of machines that already offer us coach thinking today.


Recap

To begin the journey of finding machine coworkers, start with new mindsets:

  • Identify areas of unlearning for yourself about technology and work
  • Strengthen your digital fluency: create or update your digital fluency roadmap
  • Analyze your thinking styles—what kinds of 'machine thought partners' might you want to investigate or create?
  • Browse various tech stacks and tools from innovative startups on sites like Producthunt
  • Brainstorm ideas for the work (and play!) you would do if your time were freed up by augmentation from machine coworkers
  • Explore no-code or low-code tools you can use to connect various tools to each other to create machine coworkers and workflows
  • Consider where machine coworkers could show up in your company's org chart, what their job descriptions would be and what decisions you could allow them to make without consulting with a human