Can AI be Contained?

Recently, I have been reading the history of technology especially how people from earlier times have dealt with major technical breakthroughs. My hope is that this can teach us lessons about how we deal with AI today.

From the Industrial Revolution to the digital age, each major technological shift has brought both disruption and advancement, influencing various aspects of society including labor markets, ethical norms, and regulatory frameworks.

Some of my favorite shifts, events and breakthroughs include: printing press (Gutenberg moment), Luddite movement and Internet.

In the 2020 book "Competing in the Age of AI ," Marco Iansiti and Karim Lakhani discuss the story of the Luddites as a historical parallel to current attitudes toward AI and technological change. The Luddites were early 19th-century English workers who destroyed weaving machinery as a form of protest. They believed, and for good reason, that the new technology would eliminate their jobs, making their skills obsolete and harming their livelihoods. Today we use Luddites as classic example of resistance to technology adoption. The authors use the Luddite analogy to stress that while the integration of AI into the economy can be disruptive, it also offers immense opportunities for those who can harness its potential effectively.

Similar lessons can be drawn from the invention and adoption of steam engine, railroads and electricity. It is clear that technology is not neutral and the future is not preordained. Our mindset about AI determines what happens in future.

This is the subject of a recent book on AI that I just finished reading.

Mustafa Suleyman's book, The Coming Wave: Technology, Power, and the Twenty-first Century's Greatest Dilemma expands on the concerns of AI and how we can harness and contain AI.

Since I started reading the book, two interesting developments happened:

1) Suleyman became CEO of Microsoft AI

2) TED talk on - AI becoming ...

Sidebar: AI experts belong to different schools of thought. Some are techno-optimists, others can be called techno-pessimists. Mustafa is somewhere in between. He draws upon his work experience and shares his views on difficult issues such as the potential disruption of jobs and economy. This highlights the tension that many of us in the AI field feel.

So what is this wave that this book is about? The authors define it as:

The Coming Wave: An emerging cluster of related technologies centered on AI and synthetic biology whose transformative applications will both empower humankind and present unprecedented risks.

Authors make the point that AI is unlike anything we have seen before - an invention that continues to invent.

Suleyman identifies four key features of this coming wave and discusses them in chapter 7:

• Asymmetry
• Hyper-evolution
• Omni-use
• Autonomy

Here are a few things that I found interesting:

• Highlighting the human side of the issues around AI. Not many authors take up the topic of ego of the creators, mostly we just celebrate them.
• Mentions of the role of East India company in history, which became more powerful than nations in its times and how this wave can do the same thing - this in the chapter on 'The Future of Nations'
• Conflicting trajectories for AI

Containment - Why is it needed and How do we go about it?

Containment - human agency over technology - is the key theme in the book and what differentiates it from other 'emerging tech' book. Here are a few key themes and a few points.

The Containment Problem: Technology's disposition to diffuse widely in waves and to have emergent impacts that are impossible to predict or control, including negative and unforeseen consequences.

The book dedicates a whole chapter on 10 steps toward containment, starting with safety and ending on coherence (coordinated execution).

Reading this book

If you are well-read about AI then you will want to skip the familiar stories about AlphaGo and so on. If you are new to the topic, the stories should help you frame where we stand today and how the rapid progress of AI took place.

Of course, watching the video could be enough for some!

The book is not without its shortcomings. Some of the criticism of the book from reviews on Amazon:

Heavy on Speculation, Light on Details.

Some reviewers thought this book is a political manifesto.

My take is that this book should not be viewed as a prescription on what to do -- rather viewpoints of someone who has seen AI development from close.

Note:

I plan to continue writing about history of technology and lessons learned in future posts.

Homework:

Curious about containment. Use your favorite LLM to ask about containment in AI and what should we do.