Book Review - Co-Intelligence

CO-INTELLIGENCE: Living and Working with AI. Ethan Mollick (2024, Kindle Edition)

The field of Artificial Intelligence (AI) is evolving rapidly and is generating a tremendous amount of debate and concern, albeit without a lot of clarity. To illustrate, in the context of higher education, students have access to ChatGPT and similar platforms, yet there is neither a clear understanding of its implications in relation to the teaching and learning process, administrative processes (e.g., student admissions), and other support services; nor agreement on how AI should or should not be used in each context. 

Ethan Mollick’s book, Co-intelligence, is intended to be a primer on the current state of AI for those who have limited background and understanding of the AI field. Mollick is a professor of management specializing in entrepreneurship and innovation at Wharton. He has long been involved in work on the applications of AI, especially for learning. Through Mollick’s work, he has observed that Large Language Models (LLMs), which are the new form of AI systems that power services like ChatGPT, don’t act like one would expect a computer to act. Instead, they act remarkably close to that of an “alien co-intelligence” that can interact well with humans, without being human. Mollick acknowledges that no one really knows where AI developments are heading. However, based on his experience and knowledge of the AI field, he fervently believes that we need to understand what AI is and how we can best work with AI systems in a synergistic manner, while preserving human judgment and expertise in the process. This is the genesis for his writing of this book. 

According to Mollick, the focus of Co-intelligence is on the near term, practical implications of “the early days of the AI Age”. Throughout the book, he sets the stage for a future where AI isn't just an optional tool, but rather an essential collaborative asset to be leveraged. Accordingly, the book is organized in two parts. 

Part 1 consists of three chapters. The first two chapters provide a basic introduction to what AI is (with a particular focus on generative AI systems) and how to aim for “alignment” with human interests. In the third chapter, Mollick offers four guiding principles for leveraging the potential of AI in consideration of its ethical implications and the need to preserve human identity. These principles are particularly insightful and include:

1. Always invite AI to the table: As AI proliferates, explore the implications for all aspects of your work to understand the nuances, strengths, weaknesses,  limitations, and abilities of these tools with a view to leveraging its potential as an “assistive tool“ (barring legal and ethical barriers).

2. Be the human in the loop: As AI gets more capable and requires less human help, learn how best to incorporate “collaborative oversight” by infusing human judgment and expertise in the process to ensure there are checks and balances on the accuracy and ethical use of AI-driven solutions. Accordingly, the human-in-the-loop approach fosters a sense of responsibility and accountability through maintaining control over the technology and its implications.

3. Treat AI like a person (but tell it what kind of person it is): To make the most of a collaborative relationship with AI, establish a clear and specific AI persona that best works for you, defining who the AI is and what problems it should tackle. By defining its persona, AI can take the form of “collaborative co-intelligence”.

4. Assume this is the worst AI you will ever use: There is no reason to suspect that the abilities of AI systems are going to stop growing anytime soon. Therefore, view the current generation of AI as transient and soon to be supplanted by more advanced developments. In doing so, remain open to new developments that may help you adapt to change, embrace new technologies, and remain competitive.

In Part II of the book, Mollick dives into how AI can change our lives by acting as a coworker, a teacher, an expert, and even a companion. In doing so, he illustrates the application of AI in real-world scenarios with examples from business and education. His focus is on the creative and collaborative ways that AI can be used in each context, albeit with a cautious eye on its risks. 

In the concluding chapter, Mollick draws from his personal experience and work to offer four scenarios for what may happen in the next few years in the world of AI, and what the world may look like as a result. Specifically, the scenarios he envisions include:

• Scenario 1: The current generation of AI is the best you will ever use (the least likely scenario). In this scenario, AI stagnates due to technological hurdles, government policy restrictions, among other factors. 

• Scenario 2: The growth in AI slows and advancements occur at a more measured pace. In this scenario, AI takes an increasingly larger role in our lives, but gradually enough that disruption is manageable and its benefits become more evident. 

• Scenario 3: AI continues to grow at an exponential pace. In this scenario, risks are more severe and less predictable, and current governmental systems do not have time to adjust in the usual way. Instead, AI bad actors are held in check by “good” AIs.

• Scenario 4: AI becomes as smart and capable as humans and potentially even more so. In this scenario, superintelligence emerges and human supremacy ends (for better or for worse).

Overall, Co-Intelligence is a thought-provoking exploration of the evolving relationship between humans and artificial intelligence. The book is well-written and easily digestible. Notably, Mollick acknowledges that the field of AI is rapidly evolving and that the contents of this book will become outdated quickly. That said, it serves as a useful resource, particularly for those without a technical background who are interested in learning about the current state of AI and its potential uses.