Co-Intelligence: Living and Working With AI

And this experimentation gives you the chance to become the best expert in the world in using AI for a task you know well. The reason for this stems from a fundamental truth about innovation: it is expensive for organizations and companies but cheap for individuals doing their job. Innovation comes from trial and error, which means that an organization trying to launch a new product to help a marketer write more compelling copy would need to build the product, test it on many users, and make changes many times to make something that works. A marketer, however, is writing copy all the time and can experiment with many different ways of using AI to help until they find one that succeeds.

I’m proposing a pragmatic approach: treat AI as if it were human because, in many ways, it behaves like one. This mindset, which echoes my “treat it like a person” principle of AI, can significantly improve your understanding of how and when to use AI in a practical, if not technical, sense.

AI excels at tasks that are intensely human. It can write, analyze, code, and chat. It can play the role of marketer or consultant, increasing productivity by outsourcing mundane tasks. However, it struggles with tasks that machines typically excel at, such as repeating a process consistently or performing complex calculations without assistance. AI systems also make mistakes, tell lies, and hallucinate answers, just like humans. Each system has its own idiosyncratic strengths and weaknesses, just like each human colleague does. Understanding these strengths and weaknesses requires time and experience working with a particular AI.

So how should we use AI to help generate ideas? Fortunately, the papers, and other research on innovation, have some good suggestions. When you do include AI in idea generation, you should expect that most of its ideas will be mediocre. But that’s okay—that’s where you, as a human, come into the equation. You are looking for ideas that spark inspiration and recombination, and having a long list of generated possibilities can be an easier place to start

So how do we do it? Let’s imagine that we want to come up with 20 ideas for marketing slogans for a new mail-order cheese shop. The AI can generate those for us, but we will get even better quality if we remember the principle of telling AI who it is: You are an expert at marketing. When asked to generate slogan ideas you come up with ideas that are different from each other, clever, and interesting. You use clever wordplay. You try not to repeat themes or ideas. Come up with 20 ideas for marketing slogans for a new mail-order cheese shop, make them different from each other, and make them clever and creative.

Another key aspect of idea generation is to embrace variance. Research shows that, to find good novel ideas, we likely have to come up with many bad novel ideas because most new ideas are pretty bad. Fortunately, we are good at filtering out low-quality ideas, so if we can generate novel ideas quickly and at low cost, we are more likely to generate at least some high-quality gems.

Some were assigned to use AI and some were not. The results were nothing short of astonishing. Participants who used ChatGPT saw a dramatic reduction in their time on tasks, slashing it by a whopping 37 percent. Not only did they save time, but the quality of their work also increased as judged by other humans. These improvements were not limited to specific areas; the entire time distribution shifted to faster work, and the entire grade distribution shifted to higher quality.

ChatGPT, without any specialized stock market knowledge, tended to outperform these more specialized models, working as a “powerful predictor of future stock price volatility.” In fact, it was the ability of the AI to apply more generalized knowledge of the world that made it such a good analyst, since it could put the risks discussed in conference calls into a larger context. The issue of hallucination was less important here, because the AI simply had to beat the best computer systems for accuracy, which it was able to do.

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. So now, in many ways, humanities majors can produce some of the most interesting “code.” Writers are often the best at prompting AI for written material because they are skilled at describing the effects they want prose to create (“end on an ominous note,” “make the tone increasingly frantic”). They are good editors, so they can provide instructions back to the AI (“make the second paragraph more vivid”).

Our new AIs have been trained on a huge amount of our cultural history and are using it to provide us with text and images in response to our queries. But there is no index or map to what they know and where they might be most helpful. Thus, we need people who have deep or broad knowledge of unusual fields to use AI in ways that others cannot, developing unexpected and valuable prompts and testing the limits of how they work. AI could catalyze interest in the humanities as a sought-after field of study, since the knowledge of the humanities makes AI users uniquely qualified to work with the AI.

I am only human, and in writing this book, I often found myself stuck. In previous books, that could mean a single sentence or paragraph would block hours of writing, as I used my frustration as an excuse to take a break and walk away until inspiration struck. With AI, that was no longer a problem. I would become a Cyborg and tell the AI: I am stuck on a paragraph in a section of a book about how AI can help get you unstuck. Can you help me rewrite the paragraph and finish it by giving me 10 options for the entire paragraph in various professional styles? Make the styles and approaches different from each other, making them extremely well written. In an instant, I had the paragraph written in a persuasive style, an informative style, a narrative style, and more. While I rarely used any of the text the AI produced, it gave me options and pathways forward.

We have already outsourced the worst part of writing (checking grammar) and math (long division) to machines like spell checkers and calculators, which freed us from these tedious tasks. It would be natural to use LLMs to extend the process. And this is indeed what we have seen in some early research on using AI for work. People who use AI to do tasks enjoy work more and feel they are better able to use their talents and abilities. The ability to outsource crappy, meaningless tasks to the AI can be freeing. The worst parts of your job go to AI so that you get to focus on the good stuff.

In study after study, the people who get the biggest boost from AI are those with the lowest initial ability—it turns poor performers into good performers. In writing tasks, bad writers become solid. In creativity tests, it boosts the least creative the most. And among law students, the worst legal writers turn into good ones. And in a study of early generative AI at a call center, the lowest-performing workers became 35 percent more productive, while experienced workers gained very little. In our study in BCG, we found similar effects. Those who had the weakest skills benefited the most from AI, but even the highest performers gained.

For example, we know that in-class lectures are not the most effective way to teach and that topics need to be interwoven together in order for students to retain what they know. Unhappily for students, however, research shows that both homework and tests are actually remarkably useful learning tools.

Cheating was already common in schools. One study of eleven years of college courses found that when students did their homework in 2008, it improved test grades for 86 percent of them, but it helped only 45 percent of students in 2017. Why? Because over half of students were looking up homework answers on the internet by 2017, so they never got the benefits of homework.

And while I am sure that in-class essay writing will come back in style as a stopgap measure, AI does more than help students cheat. Every school or instructor will need to think hard about what AI use is acceptable: Is asking AI to provide a draft of an outline cheating? Requesting help with a sentence that someone is stuck on? Is asking for a list of references or an explainer about a topic cheating? We need to rethink education.

By the mid-1990s, calculators were part of the curriculum and were used to complement other ways of learning math. Some tests allowed them, some did not. A practical consensus was achieved. Math education did not fall apart, though debate and research continues today, a half century after the calculator appeared in classrooms.

To some extent, AI will follow a similar path. There will be assignments where AI assistance is required and some where AI use is not allowed. In-school writing assignments on non-internet-enabled computers, combined with written exams, will ensure students learn basic writing skills. We’ll find a practical consensus that will allow AI to be integrated into the learning process without compromising the development of critical skills. 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.

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. But the moment isn’t just about preserving old types of assignments. AI provides the chance to generate new approaches to pedagogy that push students in ambitious ways.

Even slow-moving educational institutions are recognizing that teaching about AI will play an important role in education, with the US Department of Education suggesting, within just months of the release of ChatGPT, that AI will need to be embraced in classrooms.

A lot of active research is happening around the best way to “program” an LLM, but one practical implication is that it can help to give the AI explicit instructions that go step by step through what you want. One approach, called chain-of-thought prompting, gives the AI an example of how you want it to reason, before you make your request. Even more usefully, you can also provide step-by-step instructions that build on each other, making it easier to check the output of each step

If you want to do something with AI, just ask it to help you do the thing. “I want to write a novel; what do you need to know to help me?” will get you surprisingly far. And remember, AI is only going to get better at guiding us, rather than requiring us to guide it. Prompting is not going to be that important for that much longer.

This doesn’t mean we shouldn’t teach about AI in schools. It is critical to give students an understanding of the downsides of AI, and the ways it can be biased or wrong or can be used unethically.

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 main idea behind flipped classrooms is to maximize classroom time for active learning and critical thinking, while using at-home learning for content delivery. The value of flipped classrooms seems to be mixed, ultimately depending on whether they encourage active learning or not.

Imagine introducing high-quality AI tutors into the flipped classroom model. These AI-powered systems have the potential to significantly enhance the learning experience for students and make flipped classrooms even more effective. They provide personalized learning, where AI tutors can tailor instruction to each student’s unique needs while continually adjusting content based on performance. This means that students can engage with the content at home more effectively, ensuring they come to class better prepared and ready to dive into hands-on activities or discussions.

With AI tutors taking care of some of the content delivery outside of class, teachers can devote more time to fostering meaningful interactions with their students during class. They can also use insights from the AI tutors to identify areas where students might need extra support or guidance, enabling teachers to provide more personalized and effective instruction. And with AI assistance, they can design better active learning opportunities in class to make sure that learning sticks.

The issue is that in order to learn to think critically, problem-solve, understand abstract concepts, reason through novel problems, and evaluate the AI’s output, we need subject matter expertise. An expert educator, with knowledge of their students and classroom, and with pedagogical content knowledge, can evaluate an AI-written syllabus or an AI-generated quiz; a seasoned architect, with a comprehensive grasp of design principles and building codes, can evaluate the feasibility of an AI-proposed building plan; a skilled physician, with extensive knowledge of human anatomy and diseases, can scrutinize an AI-generated diagnosis or treatment plan. The closer we move to a world of Cyborgs and Centaurs in which the AI augments our work, the more we need to maintain and nurture human expertise. We need expert humans in the loop.

Instead of just iterating designs, Raj engages in a structured reflection after every project, thanks to the insights from the AI. It’s akin to having a mentor watching over his shoulder at every step, nudging him toward excellence.

It may be that working with AI is itself a form of expertise. It is possible that some people are just really good at it. They can adopt Cyborg practices better than others and have a natural (or learned) gift for working with LLM systems. For them, AI is a huge blessing that changes their place in work and society. Other people may get a small gain from these systems, but these new kings and queens of AI get orders of magnitude improvements. If this scenario is true, they would be the new stars of our AI age and would be sought out by every company and institution, the way other top performers are recruited today.

But the first society-wide benefits also begin to appear. Innovation has been slowing alarmingly in recent decades. In fact, a recent, convincing, and depressing paper found that the pace of invention is dropping in every field, from agriculture to cancer research. More researchers are required to advance the state of the art. In fact, the speed of innovation appears to be dropping by 50 percent every 13 years, slowing down economic growth.

Part of the issue appears to be a growing problem with scientific research itself: there is too much of it. The burden of knowledge is increasing, in that there is too much to know before a new scientist has enough expertise to start doing research themselves. This is also why half of all pioneering contributions in science now happen after age forty, when it used to be that younger scientists were the ones who achieved breakthroughs.

But there are already signs that AI can help. Research has successfully demonstrated that it is possible to correctly determine the most promising directions in science by analyzing past papers with AI, ideally combining human filtering with the AI software. And other work has found that AI shows considerable promise autonomously conducting scientific experiments, finding mathematical proofs, and more. It may be that the advances in AI can help us overcome the limitations of our merely human science and lead to breakthroughs in how we understand the universe and ourselves.

In some ways, however, this shift has already been occurring. In 1865 the average British man worked 124,000 hours over his lifetime, as did people in the US and Japan. By 1980, British workers spent only 69,000 hours at work, despite living longer. In the US, we went from spending 50 percent of our lives working to 20 percent. Work hours have improved more slowly since 1980. Still, UK workers now work 115 hours less a year than they did then, a decline of 6 percent. And similar changes are happening all over the world.