AI, existential threat to humanity or utopian opportunity?
Sunday, 2 June 2024
Nick Bostrom’s latest essay scene-setting is reminiscent of Plato’s dialogues – with a 21st-century twist.
There is no Athens port in the background; only a conference room where a philosopher and a few students talk about artificial intelligence and the future of humanity.
Bostrom, who is an AI philosopher, became world-famous after publishing a book on the global existential risk AI posed to the future of humankind.
His latest book, Deep Utopia, focuses on how human beings can instead harness and exploit AI’s full potential.
The Swedish-born philosopher is the founder and until recently the director of the Future of Humanity Institute at Oxford University in the UK, a research centre that was closed this year.
“What started as a small little research group, really just three people and me within the faculty of philosophy, mainly philosophers, eventually evolved to a 40-person research institute where there were hardly any philosophers left.
“It was mainly computer scientists and mathematicians and engineers and economists and political scientists,” he says.
When he published his bestelling book Superintelligence: Paths, Dangers, Strategies in 2014, AI was almost a neglected topic, Bostrom says.
“There were maybe a handful of people in the world who had started to think about ways in which future AI could pose existential risk [to our civilization].
“So it seemed pressing to try to draw attention,” he says.
The book received high praises from the likes of Elon Musk and Bill Gates, and it helped make Bostrom an influential figure around the world.
Since then he has given advice on AI to the UK parliament, the European Commission and the prime minister of Singapore, alongside having his name popping up in dystopian movies like Dream Scenario.
Ten years after he published Superintelligence, AI has become an active research field with scholars, engineers and developers trying to solve the “problem of developing scalable methods for AI alignment” – which means ensuring the development of AI was under the supervision of ethical principles and focused on given objectives.
“All the frontier AI labs now, Google DeepMind, OpenAI, Anthropic, all have research teams specifically working on this,” Bostrom says.
This month, Google released Project Astra, an AI-powered tool that allowed users to identify objects using a camera, as well as visualise and interpret writings.
In March, his latest book, Deep Utopia, came out and it is still difficult to find in New Zealand bookshops.
Deep Utopia is structured in three weaving threads: a series of six lectures given by a fictitious Nick Bostrom, a dialogue between the people attending the lectures and the assigned readings, which included letters written by a fox and the story of ThermoRex, a heater.
Bostrom says the book considered the challenges that could come up once the AI alignment problem was solved – “what happens if things really go right”.
He used to do stand-up comedy while at university and has preserved some of that irony and conciseness in his latest writings.
In Deep Utopia, a student calls the fictional character Nick Bostrom “doom-monger”, while in another part there is a sarcastic grouse about traffic jam and cars.
Bostrom describes future societies that he calls post-scarcity, post-work and post-instrumental utopias.
Western countries, where people have “overstocked fridges and around-the-clock delivery services”, optimal temperatures thanks to air-conditioning, where the spread of deadly diseases has been drastically reduced and lifespan extended, were on their way to reach the first stage of a utopian society.
However, in the future, technological advances and biomedical enhancements could transform human beings.
“For example, suppose there were a series of progressively more expensive medical treatments that each added some interval of healthy life-expectancy, or that made somebody smarter or more physically attractive,” Bostrom writes in the book.
In this “age of abundance”, a futuristic scenario where all the useful technological progress has been reached and all the inventions have been made, people would not need to work to maintain a lifestyle, but could opt to work as a leisure activity.
Such a “culture of leisure” would prioritise “enjoyment and appreciation rather than usefulness and efficiency”, while medical nanobots would work to activate endorphins to make a person feel maximal pleasure, shape the body so that there was no need to exercise daily, and edit the brain to quickly learn complex mathematical theorems.
Bostrom, who used to be a professor of philosophy at Oxford until earlier this year, says he had had enough of academic bureaucracy and administration.
He’s now looking to find a quiet and peaceful place where he can try new ideas – and if there are running trails in the wild, that would be a plus.
“I’m breathing the fresh air of freedom.”
What do you think? Email sundayletters@stuff.co.nz. Please include your full name and address.