Treat AI as a Person: Exploring Dr. Ethan Mollick’s Co-Intelligence
A couple weeks ago, a notable voice in the artificial intelligence movement, Dr. Ethan Mollick, dropped a book on artificial intelligence called Co-Intelligence.
Ethan Mollick is particularly known for his integrating AI into his business courses at the University of Pennsylvania; he was the first person I saw bring AI policies into his course syllabi in December 2022, when most folks had not even heard of ChatGPT yet.
He has also explored AI’s broader implications through his popular Substack newsletter, “One Useful Thing”, and collaborative research on AI’s impact in professional settings like consulting. Though not a computer scientist, his expertise in business innovation provides a unique, non-technical perspective on AI, making complex subjects accessible and relevant.
In the latest episode of The AI Education Conversation, I explore Co-Intelligence and offer my full review of Dr. Mollick’s book. As an appetizer for the episode, here are some of the top takeaways after reading Co-Intelligence.
A world of AI and humans may require us to treat AI as a person
While many folks I’ve talked to on The AI Education Conversation + other AI experts have asserted that assigning human like qualities to AI is problematic, comparing generative AI tools like ChatGPT to a calculator, cell phone, or computer don’t comprehensively align to the experience a person has after spending ample time exploring AI. Mollick argues we need to treat AI as a person because it is the most familiar subject on the actual strengths and weaknesses we know about — a tool with immense creative and innovative potential, but also prone to bias and persuasion. While I’m conflicted on this take, I agree with him more than I do disagree — AI is different and, while not actually human, it feels like something much deeper as you explore (Mollick argues an alien).
Relationships will change…we need to establish boundaries now
Related to the point above, Mollick also provides a fascinating overview of historical human-AI engagements and relationships since the development of chatbots in the 1950s. Even with more elementary like chatbots (i.e. Replika), humans have fostered deep emotional connections with chatbots. In a world of generative AI where these tools are engaging, can be trained to be persuasive, and have multimodal capabilities (i.e. they can talk and have a human-like avatar), we may see an explosion of human-ai intimacy as many people crave connection. How will we define and norm on acceptable human-ai interactions with this inevitability?
Curation may become a vital human lens
While we often value attributes like creativity and problem solving, Mollick describes at one point in the book how many studies point to the fact that innovation and invention have significantly decreased in the last few decades? Why? The sheer volume of information + insights proliferated daily creates a reality where it is tough to keep up on the latest research, let alone propel it further. This could present an opportunity for artificial intelligence led by humans with a strong knack for curation and synthesis, to help our society effectively navigate the information age — one where we can keep up with the latest research by analyzing macro trends, anticipating possibilities, and ramping up progress.
A vision for AI in education may still be developing
While in certain sections throughout the book, I was captivated and satisfied by the depth and texture of Dr. Mollick’s arguments, he presents some of the critical elements for what AI in k-12 education could look like in his chapter called “AI as a Tutor”. Briefly, he asserts three key points in this section, that is, we know 1:1 interactions are the most powerful way to learn, AI tools possess powerful capabilities to serve in a tutor role, and a flipped classroom model may present a model for schools to leverage AI in a way that maximizes learning. This is the chapter I was most underwhelmed by and I disagree with Mollick’s take here. While I see value in AI tools in tutoring contexts, this in of itself, will not bring about the type of transformation K-12 education needs (though he doesn’t quite claim this). Also, a flipped classroom model negates some of the harsh realities many of our schools are experiencing today — students who are struggling with motivation, tired of being placed in front of a screen, and have multiple responsibilities after the school day. A flipped classroom experience where we leverage the capabilities of strong differentiation through AI tools will likely perpetuate what is already happening in education — a system of AI haves and have nots.
Overall, I think Co-Intelligence was insightful, meaningful, engaging, and easy to read in a few days. It provides a solid foundation of key AI developments in the last two years and explores possibilities for our future world — a world with AI and humans. Ethan Mollick set out to be our guide on a journey of exploring artificial intelligence in this book and on that front, I think he succeeded.
If my exploration has encouraged you to pick up Dr. Mollick’s book, consider purchasing it here, which supports the work I do on this podcast and allows me to invest more time and resources in creating insights on AI in education.
Check out my full episode review of Co-Intelligence here.
If you have any questions about Co-Intelligence, want to share your takeaways with me after reading, or would like me to consider other texts for future book reviews, you can join the conversation at TheAIEducationConversation.com or message me on linkedin.
If you want to continue supporting the podcast, a rating, review, share on social media, or recommendation to a colleague go a long way in helping to amplify our space and host bigger conversations.
#HumansAtTheHeartOfEducation
