Co-Intelligence: Living and Working with AI by Ethan Mollick is a practical guide that explores how artificial intelligence (AI) can transform personal and professional lives when used strategically and consciously. Throughout the book, Mollick argues that AI should not be seen solely as a tool, but as a “colleague” with whom it is possible to collaborate to boost productivity, creativity, and decision-making. Through clear explanations and concrete examples, Mollick highlights how AI can improve business processes, from generating creative ideas to automating everyday tasks.
The author, however, also warns of the risks and limitations of AI, such as its tendency to “hallucinate,” that is, generate false or inaccurate information with a high degree of confidence. For effective use, he proposes four key principles: invite AI to actively participate in daily tasks; that a human person critically supervises the activity; specify the role that is expected of AI in each situation; and always assume that technology will continue to evolve, so it is very important to stay up to date.
How does AI work?
To make informed decisions, it is key to understand how modern AI works, especially large language models (LLM). These models are based on a pattern of action: the recognition of patterns through probabilistic prediction.
For pattern recognition, AI analyzes large volumes of data to identify relationships and trends and make a probabilistic prediction of the words to be placed in the speech. A simple example given by Mollick himself: if one asks AI to finish the following sentence: “I think, therefore…”; AI will always predict when asked for the word “I exist”, because it is extremely likely that this will be the case. However, if one writes something more unique like: “Martians eat bananas, because…”, the AI will offer different answers each time this instruction or “prompt” is given.
According to Mollíck, “to teach AI how to understand and generate human-like writing, it is necessary to pre-train it with a massive amount of text from various sources, such as web pages, books and other digital documents. And unlike what happened at the beginning of AI, this activity is no longer supervised.” Therefore, this author adds, “this pre-training is an interactive process, which requires very powerful computers, in a process that involves costs that exceed 100 million dollars.”
One fact that may call our attention is knowing that in this process of preparing responses, there is no real understanding on the part of the AI. The AI does not “understand” the information, it only predicts responses based on data correlations.
Ethan Mollick’s Four Principles
Here’s how to apply his four key principles in a structured and simple way.
Invite AI to the table (proactive onboarding)
Mollick recommends actively incorporating AI into everyday business tasks. Its use should be broad, unless there are legal or ethical restrictions. For example, summarizing reports: condensing financial or market reports; analyzing data: identifying sales trends and patterns; automating tasks: drafting emails, creating presentations, or scheduling meetings.
Maintaining control of the process (critical oversight)
A major risk of AI is its tendency to “hallucinate”: invent false, but compelling information. For this reason, Mollick emphasizes the importance of the human person as a critical supervisor. Some areas where oversight is essential: legal contracts: AI can generate drafts, but a lawyer must validate them; market analysis: verifying that the data and projections presented are real and based on verified sources; risk management: reviewing investment recommendations or financial predictions with human judgment.
Never blindly trust results, but use AI as an assistant, not the sole source of truth.
Treat AI like a person, and specify its role
The quality of AI results improves significantly when it is assigned a specific role when interacting with it. For example, it can be useful to ask it to act as a financial analyst: “evaluate this data from the perspective of a financial consultant”; ask it to act as an advertising campaign creator: “generate creative ideas for a product launch campaign”; ask it to act as a legal writer: “prepare a draft service contract”. By defining a role for it, AI generates responses that are more tailored to the context and needs of the company.
Assume that the version in use is the worst AI one will use (continuous improvement mentality)
Mollick emphasizes that AI is constantly evolving. What seems advanced today will soon be basic. Therefore, it is vital to maintain an attitude of continuous improvement and technological updating. That is why some of her recommendations are: continuous training: train teams in the use of AI tools and their advances; piloting and testing: deploying AI in small projects before applying it on a large scale; and avoiding complacency: not assuming that current technology is definitive or infallible.
In conclusion
AI can be a strategic ally in business management if used correctly, and a balanced approach is adopted. That is, if the advantages of AI are proactively leveraged, while still exercising human control and critical judgment.
