Lessons from Ethan Mollick

Ethan Mollick is a leading voice on AI, innovation, and entrepreneurship.

On Co-Intelligence and Working with AI

Principle 1: Always invite AI to the table. You don't know what AI can do, and its creators likely don't know everything it can do for you and your industry. Use it for everything you legally and ethically can. [1][2]
Principle 2: Be the human in the loop. AI works best with human guidance. As AI becomes more capable, your role as the human in the loop remains crucial for oversight and direction. [2][3]
Principle 3: Treat AI like a person (but tell it what kind of person it is). Interacting with AI as if it's a person, by giving it a role and context, is the easiest way to work with it, even though it is not sentient. [2][3]
Principle 4: Assume this is the worst AI you will ever use. The AI you are using now will likely be the least capable version you will ever use, as the technology is advancing at an exponential rate. [2][3]
AI has a "jagged frontier." It's surprisingly good at some tasks and unexpectedly bad at others. You can only learn this frontier by using it extensively. [4]
AI doesn't act like software, but it does act like a human being. This is a common misconception; it's more like an "alien person" than a traditional machine, which makes interacting with it by talking to it more effective. [2][3]
We have created a weird alien mind. It's not sentient but can convincingly fake it, trained on vast human knowledge and capable of creativity, yet it also fabricates information regularly. [5]
AI plausibility leads to human gullibility. Because AI-generated content often sounds correct, people can fall into a trap of not checking its work, a phenomenon he calls "falling asleep at the wheel." [1]
AI can be a powerful tool for innovation. As "connection machines," Large Language Models can generate novel concepts by finding relationships between seemingly unrelated ideas. [5]
Expertise will matter more than before. Experts are better positioned to leverage AI coworkers, fact-check its outputs, and correct its errors, ultimately getting the most out of the technology. [2]
AI levels the playing field. Studies show that humans working with AI can outperform all but the very best humans working without it, raising the performance of those who were previously struggling. [2]
There's something impossible you always wanted to do that you can now do. AI opens up new possibilities for tasks that were previously out of reach for individuals or organizations. [6]
Something you used to do that was valuable is no longer valuable. Companies and individuals need to quickly identify which of their existing skills and services are being devalued by AI. [6]
You probably shouldn't trust any video or audio recording ever again. The rise of realistic AI-generated media means that the authenticity of digital content can no longer be taken for granted. [7]
The next phase of AI development will involve more AI “agents”. These will be semi-autonomous AIs that can be given a goal and execute it with minimal human intervention. [2]

On AI in Education

Homework is over. There is virtually no homework assignment that a well-prompted AI cannot solve, which requires a fundamental shift in how we assess learning. [8]
AI cheating will remain undetectable and widespread. Instead of trying to prevent it, educators should adapt their teaching methods to incorporate AI as a tool. [9]
AI holds the potential to deliver personalized education. It can approximate the "two sigma" effect of one-on-one tutoring, providing customized learning experiences for every student. [3]
In the rush to deliver AI benefits directly to students, the role of teachers is often overlooked. AI tutors are exciting, but they do not replace the complex and essential role of a human teacher. [7]
AI can help teachers generate a ton of diverse examples. This allows for tailoring educational content to students' specific interests and needs, making learning more engaging. [7]
AI can generate different explanations and walk through ideas step-by-step. For students who are struggling, AI can provide simpler summaries to help them catch up. [7]
Just as calculators did not replace the need for learning math, AI will not replace the need for learning to write and think critically. Education will adapt to integrate AI while still teaching fundamental skills. [9]
Assignments that combine AI's capabilities with human creativity can yield transformative results. Students using AI to critique and refine their work can deepen their learning. [3]
We'll find a practical consensus that will allow AI to be integrated into the learning process. This will involve a mix of assignments where AI is required and others where its use is not allowed. [9]
With AI tutors, teachers can devote more time to fostering meaningful interactions. By handling some content delivery, AI frees up educators to focus on higher-value activities in the classroom. [9]
I have my students cheat in class. By requiring students to use AI and prompt it multiple times, the final results become unique and demonstrate their engagement with the material. [8]
AI is not just writing code; it's the brains of the operation. In educational simulations, AI can act as instructors, mentors, and observers, creating dynamic learning environments. [8]
AI has better than human hearing. Tools like Whisper can transcribe accents and mixed languages with incredible accuracy, opening new possibilities for feedback and interaction. [8]
The AI does a better job than real venture capitalists in giving feedback. When used to simulate a VC, AI can provide insightful and high-quality feedback to students on their pitches. [8]
This should be a red alert everywhere in every organization. The significant performance improvements seen with AI mean that all organizations, especially educational ones, need to urgently figure out how to integrate it. [10]

On Entrepreneurship and Innovation

Great entrepreneurs are made, not born. This is a core message of his book, "The Unicorn's Shadow," arguing that entrepreneurship can be taught and learned. [11]
Start with the resources you have today and leap into starting your business. Don't wait for the perfect idea; embrace the one that you are uniquely positioned to implement right now. [12][13]
The culture that a founder creates is one of the most durable aspects of the company. It often outlasts the founder and continues to shape the firm's traditions and success. [5][12]
Founders who use a disciplined, scientific process are more likely to succeed. Testing ideas through rigorous experimentation is a key predictor of startup success. [4][12]
There are no personality traits that predict entrepreneurial success durably. While traits like overconfidence might predict who starts a business, they don't predict who succeeds. [14]
The average age for a founder in the United States is 42. The myth of the young founder is not supported by data; experience is a significant asset. [14]
Twitter users who followed diverse groups of people were better at generating ideas. An MIT study highlights the importance of open networks for innovation. [12]
Passionate pitches are not always the most effective. Professional investors are often more convinced by a well-organized, even "boring," pitch than by sheer charisma. [15]
Amateur investors, however, are more likely to be swayed by charisma. They want to see that the founder is excited about their company. [15]
User innovators are often the source of breakthrough ideas. The people actually using a product or service are a key source of innovation for new features and offerings. [5]
Most work is not well-designed or well-thought-out. Automation and AI offer a chance to redesign jobs to be more meaningful and allow people to focus on their strengths. [16]
Boredom is not just boring; it is dangerous in its own way. It can lead to negative behaviors, and making work more engaging is a critical challenge. [5]
We have invented technologies... that boost our physical capabilities... but we have never built a generally applicable technology that can boost our intelligence. AI represents a new category of tool for humanity. [5]
Employees are already quietly experimenting with AI to improve their productivity. However, they are often hesitant to share these innovations with their employers due to fear or lack of incentives. [17]
To overcome this, leaders need to radically rethink incentives. This could include spot bonuses for impactful AI prompts or job security guarantees to encourage open innovation. [17]

On the Future and Learning

Staying Ahead of AI Advances: "Adopting a mindset of continuous learning and adaptation, rather than fixating on AI's current limits, is key to staying ahead of the curve." [18]
The World Doesn't Care What You Majored In. People often end up in careers they hadn't even heard of during their undergraduate studies, emphasizing the need for adaptable skills. [18]
You have to have a crisis when you use AI. The realization of what AI can do—how it can be more insightful or creative than you—is an unnerving but necessary experience to move past. [19]
If you're just using it like a Google or query system, it's not as powerful. Deep interaction and iteration are what unlock the true potential of AI as a partner. [20]
My goal is to bend the needle towards us thinking about practical, valuable use cases that help humans thrive. While negative outcomes are possible, the focus should be on the positive and transformative potential of AI. [16]

Learn more:

Lessons from Ethan Mollick

On Co-Intelligence and Working with AI

On AI in Education

On Entrepreneurship and Innovation

On the Future and Learning

Written by Antoine Buteau

Lessons from Jin Chen

Lessons from Marshall McLuhan