Co-Intelligence: Living and Working with AI

Who Should Read This Book and Why

• CIOs and technology leaders seeking practical frameworks for strategic AI implementation beyond technical considerations
• Cross-functional teams responsible for integrating AI into business workflows and managing organizational change

Summary

"Co-Intelligence: Living and Working with AI" is a practical guide for understanding and effectively collaborating with artificial intelligence in various aspects of life.

Published in 2024, this New York Times bestseller by Wharton professor Ethan Mollick explores how we can engage with AI as a co-worker, co-teacher, and coach.

The book examines AI's impact on business and education through real-world examples, showing what it means to think and work together with intelligent machines.

Rather than focusing on fear or hype, Mollick provides a balanced perspective that cuts through both AI evangelism and doom-mongering, offering practical frameworks for using AI's capabilities while maintaining human oversight and identity.

Key Learning Points

The Four Rules for Co-Intelligence

1. Always Invite AI to the Table Make sure AI is involved in decision-making processes
2. Be the Human in the Loop (HITL) Maintain human oversight and critical analysis when interacting with AI, as it can hallucinate, lie, and doesn't truly 'understand' information. Its primary goal is to please users, not necessarily provide accurate information
3. Treat AI Like a Person (But Choose What Type of Person) Approach AI as a human collaborator, imagining it as "an infinitely fast intern, eager to please but prone to bending the truth"
4. Assume This is the Worst AI You Will Ever Use Recognize that AI technologies will improve rapidly; today's limitations will be overcome in future iterations

The Jagged Frontier

The "Jagged Frontier" represents the unpredictable and uneven nature of AI capabilities. Based on Mollick's research, AI can excel at highly complex tasks (like creativity tests) while failing at seemingly simple ones (like basic word problems).

This concept describes how some tasks that logically seem equally difficult may actually be on different sides of AI's capability "wall" - for instance, AI might write an excellent sonnet but struggle with a precise 50-word poem.

Centaurs and Cyborgs: Two Approaches to Working with AI

Two effective approaches to navigate the Jagged Frontier:

1. Centaurs: Create a clear division of labor between human and machine, strategically switching between AI and human tasks based on their respective strengths. Like the mythological centaur with a distinct human torso and horse body.
2. Cyborgs: Blend human and machine efforts, intertwining their work and moving back and forth across the Jagged Frontier. Tasks are shared, with humans and AI collaborating on the same work, such as a human starting a sentence for AI to complete.

Task Types

Mollick categorizes tasks into three types:

• Just Me Tasks: Best handled without AI intervention
• Delegate Tasks: Assigned to AI but requiring human verification
• Automated Tasks: Fully managed by AI

AI Roles

The book explores different ways to leverage AI:

• AI as a Creative partner
• AI as a Coworker
• AI as a Coach/Tutor
• AI as a Person (with specific assigned roles)

Future Scenarios

Mollick outlines four potential future scenarios for AI development:

1. As Good as It Gets: AI development plateaus with minimal future improvements
2. Slow Growth: AI continues improving at a manageable pace, allowing society to adapt
3. Exponential Growth: Rapid advancement leading to transformative changes requiring proactive measures
4. AGI Achievement: AI potentially surpassing human intelligence

By the End of This Book You Will...

1. Understand practical frameworks for effectively collaborating with AI systems rather than merely using them as tools
2. Be able to identify which tasks should be handled by humans, which by AI, and which through collaboration
3. Know how to prompt and guide AI to achieve better results through the right role definition and guidance
4. Recognize AI's limitations and the importance of human oversight in the "human in the loop" model
5. Develop strategies for navigating the "Jagged Frontier" of AI capabilities
6. Be prepared to adapt to future AI advancements with a realistic perspective
7. Have practical approaches for implementing AI in workplace and educational settings
8. Understand the importance of experimentation in discovering effective AI applications