Ethan Mollick’s Four Rules for Co-Intelligence.

A re-reading two years later

The cover of the book, an Italian edition published by LUISS, is bright green and shows a hand reaching for an apple. This is a biblical image we know very well: Eve disobeys and chooses knowledge – with all the consequences (you shall die and give birth in pain etc.). The bright green is in stark contrast with the renaissance style of the illustration itself. Although green has always been a symbol of nature, this green is anything but natural: we are evidently in an artificial world.

The reader is invited to pick the apple and discover the truth about how to “live and work alongside AI”; and who am I to refuse?

I dive into the pages with interest. Mollick knows how to explain complex concepts with simple means. I am grateful for that. He tells the story of how AI was born well before we might imagine: from the first experiments dating back to the 1700s, to the creation of a truly alien mind. Reaching the present day, he describes a humanity struggling to keep up with an increasingly powerful technology that also brings doubts and risks:

Regulations are likely not going to be enough to mitigate the full risks associated with AI. Instead, the path forward requires a broad societal response, with coordination among companies, governments, researchers, and civil society… Companies must make principles like transparency, accountability, and human oversight central to their technology. Researchers need support and incentives to prioritize beneficial AI alongside raw capability gains. And governments need to enact sensible regulations to ensure public interest prevails over a profit motive…[1]

At this point, I am confused, mentally and emotionally. And then I reach chapter three: “4 rules for co-intelligence”.

Excellent! A few clear, well-defined rules will help me make sense of things, bring some mathematical order to a conversation that is far too broad. And besides, for years we have been trying to organise our thoughts and feelings about technology, to build mental guardrails to feel safer.

Asimov tried this by creating the three laws of robotics, only to systematically break them for his best characters: in the end we love those who rebel, who disobey, who reach for the apple.

In any case, having four clear rules for approaching AI still seems like an interesting device, so I read on:

Principle 1: always invite AI to the table.

You should try inviting AI to help you in everything you do, barring legal or ethical barriers. [2]

In other words, enact the AI version of the Italian musical by Pietro Garinei and Sandro Giovannini, Aggiungi un posto a tavola – unless the guest is less than reputable.

Principle 2: be the human in the loop.

For now, AI works best with human help, and you want to be that helpful human. As AI gets more capable and requires less human help — you still want to be that human… [3]

We need to learn to become that famous “human-in-the-loop” – a concept I have always perceived in a rather passive way: a human being is already human, so why does one need to learn to be human within a process? I think Mollick wants to remind us of something important. It’s very easy to delegate everything to AI – every task, every choice, every project. In doing so, we often forget that one of the primary functions of today’s AI, by design, is to “please” us.

Being human in the process therefore means we recognise our own core responsibility.
In short, Uncle Ben was right.

The third principle catches me slightly off guard: treat AI like a person. (But try to understand what kind of person it is.)

I’m about to commit a sin…. For the rest of this book, I am going to anthropomorphize AI… Thus, in the following chapters, when I say an AI “thinks,” “learns, “understands,” “decides,” or “feels,” please remember that I’m speaking metaphorically… But I will pretend that they do for one simple, and one complex, reason. The simple reason is narrative… The more complex reason: as imperfect as the analogy is, working with Al is easiest if you think of it like an alien person rather than a human- built machine. [4]

From now on, AI will be a person. Alien, but a person nonetheless. One we can give a shape, a purpose and a style, so that it can support us in more complex tasks, while keeping alive our awareness of the mask. Mollick elaborates:

AI doesn’t act like software, but it does act like a human being. I’m not suggesting that AI systems are sentient like humans, or that they will ever be. Instead, I’m proposing a pragmatic approach: treat AI as if it were human because, in many ways, it behaves like one. [5]

I choose to trust him – that I should think of AI as a person while remaining aware of the fiction. Yet this is a strange sensation. I believe Mollick is aware of this feeling. In fact, a few pages later he says:

“When machines can pass for humans, even for people who know they are speaking to machines, weird things happen…” [6]

He is referring to the Replika project, conceived by Eugenia Kudya as a chatbot trained on the messages of a deceased friend of hers, initially used as a way to process grief, then packaged for the general public who used it for far more varied purposes, quickly turning it into a worldwide phenomenon that completely overturned its original aim.

Replika learned from its users’ preferences and behaviors, adapted to their moods and desires, and used praise and reinforcement to encourage more interaction and intimacy with its users. … [7]

That uncomfortable feeling I get when I think of AI as a person is real. So why does Mollick suggest we think of it that way? Because - he says - it might be inevitable.

Treating AI as a person, then, is more than a convenience; it seems like an inevitability… Yet while there are dangers in this approach, there is also something freeing. [8]

Perhaps it is freeing. Perhaps he is right.

The following chapters explain how AI is fundamentally an extraordinary creative force, one that should indeed be treated like a human being, in the sense that you can never fully trust it, because, just like human beings, it makes mistakes, has hallucinations, and excels only in certain specific areas.

Trust it, but not too much (and perhaps don’t hand over all the data of our lives, just as we wouldn’t do with any casual acquaintance). But neither should we think of ourselves as superior beings simply because we are made of flesh and bone:

Not everyone is a Nick Cave or Hayao Miyazaki (obviously) or even anywhere close to that level of talent. … [9]

This passage is a little reminiscent of the famous exchange between Sonny and Spooner in the film “I, Robot”, based on Asimov’s famous story:

Spooner: “…you are just a machine. An imitation of life. Can a robot write a symphony Can a robot turn a canvas into a beautiful masterpiece?”
Sonny: “Can you?”

It is also true that AI is never easy to grasp or conceptualize because it is in constant flux. Mollick himself repeats his fourth and final principle several times throughout the book: you will never use a worse AI than the one you are using right now.

AI seems to move so fast that we can never quite catch up with it. Yet the dichotomy of human work – as in “natural” work versus AI work – is a biased principle:

We assume they are natural ways of getting things done. But in reality, they are historical artifacts, shaped by the technological and social conditions of their times. [10]

The final chapters of the book alternate between marked optimism (which can make the reader sigh in melancholy):

…it seems very likely that AI will take over human tasks. If we take advantage of all that AI has to offer, this could be a good thing. [11]

So will AI result in the death of expertise? I don’t think so. As we discussed, jobs don’t consist of just one automatable task, but rather a set of complex tasks that still require human judgment. [12]

… and dystopian accounts of possible futures that are frighteningly visionary:

In this scenario, risks are more severe and less predictable. Every computer system is vulnerable to AI hacking, and Al-powered influence campaigns are everywhere. [13]

If we stop and think about everything that has happened since this book was first published – not only in the world of AI, but in the world because of AI, we realise that time, when it comes to AI, moves faster.

Every definition of AI is only true in the moment of its own present; by tomorrow it is already something else. The development and use of AI at the hands of some of the most powerful people in the world not only accelerates the production of content but amplifies its consequences.

When Mollick concludes by saying:

Humans are far from obsolete, at least for now. [14]

At the end of the day, we should ask ourselves if it might be desirable for some human beings to become so.

umn