AI 2041: Ten Visions for Our Future by Kai-Fu Lee and Chen Qiufan
480 pages, Hardcover
Published, 2021 by Crown Currency
Since I
interviewed Stanley Chen Qifan during Global Time Slip, I intended to read this
first, but my library hold didn’t come in time. That interview was very cool,
by the way. The set-up for this book is incredibly smart, and when it was
released in 2021, it should have been a bigger topic. I am not sure if the
world coming out of the pandemic was ready for what revolution was
coming.
Stanley
seemed sensitive about how the world has changed in the six years since they
wrote this, but I still think the book is important. Most commentators
talking about AI only reference China in the context of “We have to beat
them.” So a book-length conversation between the former president of
Google China, Kai-Fu Lee, and the author of the brilliant SF novel The Waste
Tide is very important work indeed, even if I rolled my eyes at many of
Lee’s pro-AI stances.
It is a
long book, and many of the stories have different translators. Ten Visions of
the Future includes a story and a non-fiction essay on each topic. Qifan uses
the story often to warn about the dangers, and Lee balances it with optimistic
takes. This whole book feels like a season of Black Mirror. The push-pull of
the authors is actually helpful, even if I land on the skeptical side.
Kai-Fu Lee
and Chen Qifun worked together at Google China, despite having very different
opinions, at least that is the feeling I get reading the stories. It is not
like one is a Luddite and the other is pro-AI; the suggestions are plus and
minus for most technology as laid out here. It is Qifan’s job to speculate, so
by nature he leans toward warning.
As a short
story collection, the 10 stories are not interconnected; they take place in
various versions of the year 2041. They are set in various places around the
globe, which alone is an exciting aspect of this book. Different settings in
Asia, Australia, South America, and Africa help this book feel global.
Pretty
much every page had important points to make. I pulled quotes from the book
that were on pages I dog-eared, and it is a mix of fiction and essays. Things
that stood out were often elements of technology that made me nervous. The
first story that really hooked me was about a Deepfake hacker in Nigeria. It
proposes the idea of deep fake masks that could be worn.
“The
more power of Deepmask he excavated, the more his addiction for the mask grew.
It concealed his real face, so that he was able to let his feelings pour out
and run free, without exposing himself to danger or shame.”
It is
interesting for a Chinese SF writer to write about deep fake masks that can
fool facial recognition. Facial recognition has been an important tool of the
dictatorship in China. It is a really important tech to explore. In my limited
experience talking with Chinese scholars of the genre (I have met a few), they
reject the notion that they have it worse than us for internet privacy.
The setting in Nigeria is interesting. Legos is trying it sell itself as a city
that could be the next Dubai. It is a good first example of the way the stories
will tackle global and technological issues.
The
nonfiction parts of the book are very educational, and I could’ve read a whole
book of them. I say that despite most of the essays being disturbingly pro-machine.
“Agriculture is surprisingly low-hanging fruit. While manufacturing a phone,
a shirt, or a shoe is completely different, fertilizing spray insecticide, and
seeding are relatively similar for many types of crops. Drones can already do
these three tasks for many types of crops, while robots are harvesting apples,
lettuce, and other fruits and vegetables today. Robotics will reduce the cost
of agriculture in time, offering the promise of reducing food insecurity around
the world as well.”
As you can
see, Kai-Fu Lee doesn’t consider the farm worker or the effect on the wider
society. For example, he talks of education, and implies that children could
hold their school in their hands. Only considering how a child has information
dispensed. Not anything about the social implications of not having a school,
or teachers, would have on the children or society.
One of the
best predictions the book made came in geo-political commentary in one of the
stories…
“The
destruction had been unprecedented. Terrorist attacks had been staged at the
seven major oil routes of the world. More than 60 million barrels of crude were
transported from major production areas to the rest of the world every day, and
most of it passed through a handful of narrow waterways: the Strait of Hormuz,
the Strait of Malacca, the Suez Canal, the Danish straits, the Bob L Mendez St.
the Turkish straits, and the Panama now.
Choking
these throats was like cutting off oxygen to a human body.”
I was
reading this as Agent Orange’s stupid war in Iran was just starting, and that
last part seemed all too obvious to Qifan writing in 2020.
The most
disturbing thing about the book, in my opinion, came in this defense of the
creation of these alien minds that have never lived. Minds that don’t have to
eat, sleep, exercise, or just flat-out live scare me. Lee seems certain there
will always be space for humans.
“AI’s
mind is different from the human mind. In twenty years, deep learning and its
extensions will beat humans on an ever-increasing number of tasks, but there
will still be many existing tasks that humans can handle much better than deep
learning. There will even be some new tasks that showcase human superiority,
especially if AI’s progress inspires us to improve and evolve. What’s important
is that we develop useful applications suitable for AI and seek to find
human-AI symbiosis, rather than obsess about whether or when deep-learning AI
will become AGI.”
Human-AI
Symbiosis is happening, AI agents renting humans from a website for Rent a
human is already paying folks to do jobs in meatspace. Lee thinks it will
inspire us to evolve and progress, but I think he is wrong. It is creating a system
where hypercharged Capitalism drains the workforce of white-collar and heavy
labor jobs alike. We might need humans to be plumbers or teachers now but this
book loves to think about ways the machines can “improve” our lives by making
us useless.
“Note
from Kai Fu: AI and other technologies will drive down the cost of all goods,
most of which will be produced for next to nothing. For the first time in human
history, developed countries could eradicate poverty and hunger. If this
happens, would money be phased out? If so, what would take money's place to
motivate people to live purpose filled lives? where does any economic theory
apply anymore?”
These Tech
Bros are not reading Marx or Kropketin, their theories are more money for them.
I think about the AI in Brunner’s 1969 The Jagged Orbit. The computer
had maximized the weapons company it managed so well that they killed off all
their customers. It had to go back in time to warn itself. The thing is, we
don’t understand these alien minds' motivations. Why would Gemini tell Jonathan
Galavans to kill himself? Why did Bing’s AI threaten to blackmail
employees?
I don’t
foresee these non-human intelligences understanding what hunger and poverty
actually mean. I think AI 2041 is a great book for asking questions. When it
tries to give answers is when I am skeptical. Still, it is a book worth
reading.
PS:
Also, it
should be noted that the only reference to PKD was in a story about Crypto… “It
was disguised in a limited edition artwork called “Does Hal Dream of Encrypted
Gold?” Only those deeply familiar with Bitcoin history would get the Philip K
deck reference, how it didn't refer to the killing machine Hal 9000 in 2001: A
Space Odyssey, but to the earliest implementation of the reusable proofs of
work system, the man who received the first Bitcoin transfer…”
PSS:
Things
AI 2041generally predicted correctly.
Deepfake explosion — Predicted massive growth and most numbers look like
a 900% annual growth. AI in Education, while we don’t have handheld
schools, the reality is 57% of universities now prioritize and accept AI usage.
Voice cloning is a reality: seconds of audio now create convincing clones.
Insurance AI, predicted deep learning would transform the insurance
industry, and while they are not tracking your heart rate live, they are affecting
pricing.
Things AI
2041generally got wrong. Lee’s Artificial Generalized Intelligence
timeline was decades out, and the tech bro D bags think anytime between now and
2028. Self-driving cars are here, but their adoption is slower than they
thought. I think it is fair to say the book