Co-Intelligence: Living and Working with AI

by Ethan Mollick

A study by researchers at the University of Chicago used ChatGPT to analyze the conference-call transcripts of large companies, asking the AI to summarize the risks that companies faced.

ChatGPT, without any specialized stock market knowledge, tended to outperform these more specialized models, working as a “powerful predictor of future stock price volatility.”

But what happens when AI touches on the most deeply human creative tasks—art?

The meaning of art is an old debate and one unlikely to be resolved in this book or any other.

AI is trained on vast swaths of humanity’s cultural heritage, so it can often best be wielded by people who have a knowledge of that heritage. To get the AI to do unique things, you need to understand parts of the culture more deeply than everyone else using the same AI systems.

Not everyone is a Nick Cave or Hayao Miyazaki (obviously) or even anywhere close to that level of talent. But many people want to express themselves creatively, and remarkably few feel they can.

Generative AI is giving people new modes of expression and new languages for their creative impulses—sometimes literally.

if you work interactively with the AI, the outcome doesn’t feel generic, it feels like a human did it. That said, it would be naive to see only the upside here. Especially as AI work becomes easy to generate at the push of a button. I mean that literally, as every major office application and email client will include a button to help you create a draft of your work. It deserves capital letters: The Button.

Everyone is going to use The Button.

The implications of having AI write our first drafts (even if we do the work ourselves, which is not a given) are huge. One consequence is that we could lose our creativity and originality.

Another consequence is that we could reduce the quality and depth of our thinking and reasoning.

There is already evidence that this is going to be a problem. The MIT study mentioned earlier found that ChatGPT mostly serves as a substitute for human effort, not a complement to our skills.

But AI will make a lot of previously useful tasks meaningless.

We are going to need to reconstruct meaning, in art and in the rituals of creative work.

almost all of our jobs will overlap with the capabilities of AI.

AI has the potential to automate mundane tasks, freeing us for work that requires uniquely human traits such as creativity and critical thinking—or, possibly, managing and curating the AI’s creative output, as we discussed in the last chapter.

There is danger in working with AIs—danger that we make ourselves redundant, of course, but also danger that we trust AIs for work too much.

When the AI is very good, humans have no reason to work hard and pay attention. They let the AI take over instead of using it as a tool, which can hurt human learning, skill development, and productivity. He called this “falling asleep at the wheel.”

AI is currently terrible at jokes, unless you absolutely love dad humor.

It does a good job approximating empathy, creativity, and humanity. Trying to find things that AI can definitely not do because they are uniquely human may ultimately be challenging.

Almost all the writing in this book is a Just Me Task. There are three reasons for this. First, the AI is good at writing, but not that good at writing with a personal style.

Delegating the task to an AI, no matter how sophisticated, could risk losing that personal touch, and the process of writing helps us think.

The third reason I won’t delegate my writing to an AI is the delicate issue of copyrights and the law.

Today, AI makes too many mistakes to use in an automated fashion. Though that starts to change when other systems enforce the accuracy of AI answers.

Centaur work has a clear line between person and machine,

Cyborgs blend machine and person, integrating the two deeply.

Using AI stopped me from ever losing momentum, and it often gives me ideas I never could have come up with before. Using AI as a co-intelligence, as I did while writing, is where AI is the most valuable.

shadow IT use is common in organizations, but it incentivizes workers to keep quiet about their innovations and productivity gains.

All this shadow use leads to the final concern, the justified worry that workers might be training their own replacements by figuring out how to work with AI. If someone has figured out how to automate 90 percent of a particular job, and they tell their boss, will the company fire 90 percent of their coworkers? Better not to speak up.

companies that commit to maintaining their workforce will likely have employees as partners who are happy to teach others about the uses of AI at work, rather than scared workers who hide their AI for fear of being replaced. Convincing employees that this is the case is another matter.

managers and leaders must decide whether and how to commit themselves to reorganizing work around AI in ways that help, rather than hurt, their human workers.

We often take for granted the systems we use to structure and coordinate work in our organizations. We assume they are natural ways of getting things done. But in reality, they are historical artifacts, shaped by the technological and social conditions of their times. The organizational chart, for example, was originally made to run railroads in the 1850s. Developed by early railroad barons, they created a hierarchical system of authority, responsibility, and communication that enabled them to control and monitor the operations of their railroad empire. Enabled by the telegraph, they integrated humans into a clear hierarchy, with bosses giving orders that flowed through the rails and telegraph lines to the workers at the bottom of the chart. The system was so successful that it was soon adopted by other industries and organizations, becoming the standard model of bureaucracy for the twentieth century.

By acting as a co-intelligence managing work, or at least helping managers manage work, the enhanced capabilities of LLMs could radically change the experience of work. A single AI can talk to hundreds of workers, offering advice and monitoring performance. They could mentor, or they could manipulate. They could guide decisions in ways that are subtle or overt.

But, unlike the cold, impersonal algorithm of Lyft or Uber, LLMs might also provide feedback and coaching to help workers improve their skills and productivity. AI’s ability to act as a friendly adviser could sand down the edges of algorithmic control, covering the Skinner box in bright wrapping paper.

Written by AI?

Humans are built deep into the fabric of every aspect of our organizations. You cannot easily replace a human with a machine without tearing that fabric. Even if you could replace a doctor with an AI overnight, would patients be okay being seen by a machine?

while jobs usually adjust to automation, they do not always, at least not for everyone.

Of course, there are also reasons why AI might be different from other technological waves. It is the first wave of automation that broadly affects the highest-paid professional workers. Plus, AI adoption is happening much more quickly, and much more broadly, than previous waves of technology. And we are still unclear as to what the limits, and possibilities, of this new technology are, how quickly they will continue to grow, and how ahistorical and strange the effects might be.

as Amara’s Law, named after futurist Roy Amara, says: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”

Here’s a secret: we have long known how to supercharge education; we just can’t quite pull it off. Benjamin Bloom, an educational psychologist, published a paper in 1984 called “The 2 Sigma Problem.” In this paper, Bloom reported that the average student tutored one-to-one performed two standard deviations better than students educated in a conventional classroom environment. This means that the average tutored student scored higher than 98 percent of the students in the control group

This is where AI comes in. Or where AI, hopefully, comes in. As remarkable as today’s AIs are, we have not reached the point where they can replace human teachers with magical textbooks. Though we are certainly at an inflection point where AI will reshape how we teach and learn, both in schools and after we leave them. At the same time, the ways in which AI will impact education in the near future are likely to be counterintuitive. They won’t replace teachers but will make classrooms more necessary. They may force us to learn more facts, not fewer, in school. And they will destroy the way we teach before they improve it.

Unhappily for students, however, research shows that both homework and tests are actually remarkably useful learning tools. So it is a blow that the first impact of Large Language Models at scale was to usher in the Homework Apocalypse.

Just as calculators did not replace the need for learning math, AI will not replace the need for learning to write and think critically. It may take a while to sort it out, but we will do so. In fact, we must do so—it’s too late to put the genie back in the bottle.

So students will cheat with AI. But as we saw with user innovation earlier, they also will begin to integrate AI into everything they do, raising new questions for educators.

The Homework Apocalypse threatens a lot of good, useful types of assignments, many of which have been used in schools for centuries. We will need to adjust quickly to preserve what we are in danger of losing and to accommodate the changes AI will bring. That will take immediate effort from instructors and education leaders and clearly articulated policies around AI use.

Being “good at prompting” is a temporary state of affairs.

Prompting is not going to be that important for that much longer.

AI tutoring will likely become excellent, but not a replacement for school. Classrooms provide so much more: opportunities to practice learned skills, collaborate on problem-solving, socialize, and receive support from instructors. School will continue to add value, even with excellent AI tutors. But those tutors will change education. They already have. Just a few months after the release of ChatGPT, I noticed that students were raising their hands less to ask basic questions.

In the longer term, however, the lecture is in danger.

One solution to incorporating more active learning is by “flipping” classrooms. Students would learn new concepts at home, typically through videos or other digital resources, and then apply what they’ve learned in the classroom through collaborative activities, discussions, or problem-solving exercises.

The biggest danger to our educational system posed by AI is not its destruction of homework, but rather its undermining of the hidden system of apprenticeship that comes after formal education.