32 Quotes from Ethan Mollick's New Book: Co-Intelligence

Practical advice from an GenAI expert

Ethan Mollick is one of the leading voices teaching us how to use GenAI for practical use in work and education.  Like many, I am a big admirer and reader of Ethan Mollick’s research and writing, and was delighted to read his wonderful new book, Co-Intelligence: Living and Working with AI.

The book is divided into 2 parts: Part 1 - Describing AI and Part 2 - Practical Applications.

Key concepts

4 key principles

• Principle 1: Always invite AI to the Table

• Principle 2: Be the human in the loop

• Principle 3: Treat AI like a person (but tell it what kind of person it is)

• Principle 4: Assume this is the worst AI you will ever use

“The Button”- start tasks with AI or “the button”

3 types of tasks

• Just Me Tasks

• Delegate Tasks

• Automated Tasks

4 future scenarios for AI

• Scenario 1: As Good As it Gets 

• Scenario 2: Slow Growth 

• Scenario 3: Exponential Growth

• Scenario 4: The Machine God

Key Quotes (bold is my emphasis)

Intro

• I discovered something remarkably close to an alien co-intelligence (xv)

• Where previous technical revolutions often targeted more mechanical and repetitive work, AI works, in many ways, as a co-intelligence….[AI] can lead to 20-80% improvement in productivity across a wide variety of tasks….in contrast, when steam power, the most fundamental General Purpose Technology improved productivity 18-22% (xvii)

• AI has blown through both the Turning Test and the Lovelace Test…it is not entirely clear why AI can do all these things….no one knows where this is headed” (xviii)

Chapter 1: Creating Alien Minds

• The crazy thing is that no one is entirely sure why a token prediction system resulted in an AI with such seemingly extraordinary abilities (24)

• Where AI works bests, and where it fails, can be hard to know in advance (25)

Chapter 2: Aligning the Alien

• Current tools launched by small teams and used by amateurs…..it is likely that national defense organizations in a dozen countries are spinning up their own LLMs, ones without guardrails (42)….This could include consensual intimate imagery, political disinformation targeting public figures, or hoaxes aimed at manipulating stock prices… allow nearly anyone to generate convincing fakes undermining privacy, security, and truth.   And it is definitely going to happen (42)

Chapter 3: Four Rules for Co-intelligence

• Principle 1: Always invite AI to the Table

• Principle 2: Be the human in the loop

• Principle 3: Treat AI like a person (but tell it what kind of person it is)… As imperfect as the analogy is, working with AI is easiest if you think of it like an alien person rather than a human-built machine (57)…..You must establish a clear and specific AI personal, defining who the AI is and what problems it should tackle (57)

• Principle 4: Assume this is the worst AI you will ever use…The next phase of AI development will involve more AI “agents” – semi-autonomous Ais that can be given a goal (“plan a vacation for me”) and execute it with minimal human help (61)….There is no reason to suspect the abilities of AI systems are going to stop growing anytime soon (62)

• Many things that once seemed exclusively human will be able to be done by AI.   So, by embracing this principle, you can view AI’s limitations as transient, and remaining open to new developments will help you to adapt to change…and remain competitive in a fast-paced business world (62)

Chapter 4 AI as a Person

• A common misconception tends to hinder our understanding…the belief that AI, being made of software, should behave like other software….AI is terrible at behaving like traditional software (65)

• AI doesn’t act like software, but it does act like a human being.  I’m not suggesting AI systems are sentient like humans…Instead, I’m proposing a pragmatic approach: treat AI as if were human because, in many ways, it behaves like one….”treat it like a person” principle of AI can significantly improve your understanding of how and when to use AI in a practical, if not technical, sense….AI excels at tasks that are intensely human (66)

• AI struggles with tasks that machines typically excel at (67)

• Though it has no morality on its own, it can interpret our moral instructions (69)

• It can be extremely challenging, in the heat of the conversation, to remember that you’re not talking to a living, conscious person (85)`

• Treating AI as a person, then is more than convenience, it seems like an inevitability (91)

Chapter 5 AI as a Creative

• AI cannot be easily used for mission-critical tasks that require precision or accuracy (98)

• This is the paradox of AI creativity.  The same feature that makes LLMs unreliable and dangerous for factual work also makes them useful.  The real question become how to use of its strengths while avoiding its weaknesses.(99)

• Are these [things created by AI] original ideas?  It is often very hard to tell.  The AI is not explicitly searching a database of ideas; instead, it relies on its training to find connection (102)

• AI should be invited to any brainstorming session you hold (106)

• When faced with the tyranny of the blank page, people are going to push The Button.  It is so much easier to start with something than nothing… the implications of having AI write our first drafts are huge (119)

Chapter 6 AI as a coworker

Types of tasks

• Just Me Tasks. They are tasks in which the AI is not useful and only gets in way, at least for now…we may reserve Just Me Tasks for personal or ethical reasons (130-1)

• Delegate Tasks.  These are tasks you assign the AI and may carefully check, but ultimately do not want to spend a lot of time on (133)

• Automated Tasks.  Ones you leave completely to the AI and don’t even check on…This is likely to be a very small category…for now (135)

• In study after study, the people who get the biggest boost from AI are those with the lowest initial ability – it turns poor performers into good performers….AI acts as a great leveler, turning everyone into an excellent worker. (157)

8 AI as a Coach

• I have been making the argument that expertise is going to matter more than before, because experts may be able to get the most out of AI coworkers and are likely to be able to fact-check and correct AI errors (187)

• We already know one major effect of AI:  it levels the play field….in field after field, we are finding that humans working with an AI co-intelligence outperforms all but the best humans working without AI. (188-9)

Chapter 9 AI as of Future

4 Scenarios

• Scenario 1: As Good As it Gets.  

• Scenario 2: Slow Growth.  …Most exponential growth slows..AI hits a barrier (199). AI comes to take a larger and larger role in our lives but gradually enough that the disruption is manageable (203)

• Scenario 3: Exponential Growth…..AI companies might use AI systems to help create the next generation of AI software.  Once this process starts, it may be hard to stop.  And at this pace, AI becomes hundreds of times more capable in the next decade (203-4)….In this scenario, risks are more severe and less predictable (204)…AIs a hundred times better than GPT-4 start to actually take over human work …it will require a major rethinking of how we approach work and society… shortened work weeks, universal basic income, …

• Scenario 4: The Machine God….Machines reach AGI and some form of sentience.  They become as smart and capable as humans. Yet there is no particular reason that human intelligence should be the upper limit (207)…..There is no theoretical reason why this can’t happen, but also no reason to suspect that it might (208)

---

Notable upcoming events

• Apr 12 (virtual, free). Weekly GenAI Learning Lab (Zoom call). Sign up via Meetup

• Apr 18 (all day - in-person (there is no virtual option); free): Forging the Future of Business with AI at MIT Media Lab. SOLD OUT A stunningly impressive set of 65 speakers is confirmed.

• May 3 (virtual, free). AI Blueprint for MA monthly Zoom call. AI Blueprint for MA is a volunteer effort to help attract and retain AI talent to the State of MA in USA. Email Paul Baier to be added to the calendar event. This event is open to the public.

  

• May 16 (in person, paid) Leveraging AI Conference in Sunnyvale, CA.
  The event is geared toward professionals working in medium and large organizations who are involved in assessing the AI landscape; setting up pilot tests; scaling what works; and developing short-, medium-, and long-term strategies. GAI Insights Co-Founder John Sviokla will be among the speakers, as will leaders from Amazon, Goodyear, Airbus, Robust.ai, Mastercard, and more. GAI Insights community will save $100 by using the promo code GAI24 here.

  

• June 4 (in person, paid). The AI Advantage in Sales workshop presented by the Distribution Strategy Group. This 3 hours, hands-on workshop shows you how to use GenAI to increase your sales effectiveness to drive cash to your pocket.

• Oct 7-8 (2 days. in-person, paid): GAI Insights’ Generative AI World 2024 in Boston, MA. Currently accepting Call for Speakers

---

“Care about people’s approval and you will be their prisoner” — attributed to Lao Tzu

Onward,
Paul