For months, social and news media have been flooded with alarming headlines warning of the extinction of the human race at the hands of hypothetical rogue AIs. Much of this content can be dismissed as clickbait, and some of it is deeply silly, but at the core of the argument are some very serious and well-respected individuals. Chief among them are Yoshua Bengio and Geoffrey Hinton, two of the three “Godfathers of AI”, whose foundational work in machine learning techniques has enabled the current boom in AI capabilities. Both now feel that these need to be carefully regulated to prevent a possible future apocalypse.
The voices on the opposite side of the argument have been led by the third “Godfather”, Yann LeCun, but being rather more mundane, they have received substantially less publicity. That changed last week when LeCun was joined by another big beast of the tech world, Marc Andreessen, who published a remarkable 7000 word essay entitled “Why AI Will Save the World”. Is he right? Will AI take on the mantle of superhero, and solve all the planet’s ills? And what’s really going on behind this heated argument?
EVERY AI, EVERYWHERE, ALL AT ONCE
Andreessen’s case for AI as superhero is wide ranging, but much of it focuses on the personal and professional benefits that we will derive as individuals from AI’s influence in our lives. He describes a future where every child has an AI tutor that is “infinitely patient, infinitely compassionate, infinitely knowledgeable, infinitely helpful,” which then becomes an equally infinite “assistant/coach/mentor/trainer/advisor/therapist” when we pass into adulthood, with an emphasis on the professional benefits of constant AI mentoring.
This story is part of our weekly briefing. Sign up to receive the FREE briefing to your inbox.
For months, social and news media have been flooded with alarming headlines warning of the extinction of the human race at the hands of hypothetical rogue AIs. Much of this content can be dismissed as clickbait, and some of it is deeply silly, but at the core of the argument are some very serious and well-respected individuals. Chief among them are Yoshua Bengio and Geoffrey Hinton, two of the three “Godfathers of AI”, whose foundational work in machine learning techniques has enabled the current boom in AI capabilities. Both now feel that these need to be carefully regulated to prevent a possible future apocalypse.
The voices on the opposite side of the argument have been led by the third “Godfather”, Yann LeCun, but being rather more mundane, they have received substantially less publicity. That changed last week when LeCun was joined by another big beast of the tech world, Marc Andreessen, who published a remarkable 7000 word essay entitled “Why AI Will Save the World”. Is he right? Will AI take on the mantle of superhero, and solve all the planet’s ills? And what’s really going on behind this heated argument?
EVERY AI, EVERYWHERE, ALL AT ONCE
Andreessen’s case for AI as superhero is wide ranging, but much of it focuses on the personal and professional benefits that we will derive as individuals from AI’s influence in our lives. He describes a future where every child has an AI tutor that is “infinitely patient, infinitely compassionate, infinitely knowledgeable, infinitely helpful,” which then becomes an equally infinite “assistant/coach/mentor/trainer/advisor/therapist” when we pass into adulthood, with an emphasis on the professional benefits of constant AI mentoring.
Science and politics are singled out as two areas where this will be especially useful, leading to cures for disease and a solution to climate change, and Andreesen even goes so far as to claim AI will make war better. Wartime deaths will be drastically reduced, he claims, as “every war is characterized by terrible decisions made under intense pressure and with sharply limited information by very limited human leaders”. Quite how war occurs in this infinitely enlightened future society is not explained, but Andreessen closes his argument with a warning that the real danger of restricting AI development is not existential catastrophe, but of the West falling behind China, who he alleges are developing their own AI systems for solely autocratic purposes.
BOOTLEGGERS AND BAPTISTS
China is not the only target of Andreessen’s ire; he also reserves harsh words for those calling for AI’s regulation, labeling them as “bootleggers and baptists”. This is a reference to the prohibition era in the US, where two groups were calling for a ban on alcohol sales; not just the true believer baptists, who sincerely believed in the moral rectitude of their campaign, but the bootleggers, who supported it to drive up the value of their illicit alcohol. Andreesen labels the baptists as “cultists”, then goes on to point fingers at several groups he describes as “bootleggers masquerading as baptists”, depicting their claims as an attempt to fortify their own profits from early AI research.
Midjourney's take on "Bootleggers and Baptists".
Andreessen’s swipe at perceived bootleggers is wrapped in a broader vision of AI’s economic impact, which he expects will be enormous. In a lengthy section - referencing the unlikely bedfellows of Adam Smith, Karl Marx and Milton Friedman - he lays out a case for technology as a job creator, not a job destroyer, but also argues that if the doomer predictions were correct, and AI replaced all human labor, this would lead to enormous economic benefits across human society: “It would mean a takeoff rate of economic productivity growth that would be absolutely stratospheric, far beyond any historical precedent. Prices of existing goods and services would drop across the board to virtually zero. Consumer welfare would skyrocket. Consumer spending power would skyrocket. New demand in the economy would explode. Entrepreneurs would create dizzying arrays of new industries, products, and services, and employ as many people and AI as they could as fast as possible to meet all the new demand.”
A META META-ARGUMENT
These are bold claims, and they go much further than those made by Yann LeCun, who has restricted himself to stating that the current generation of AIs are fundamentally incapable of incredible feats of superhero/supervillain behaviors. LeCun’s position is far simpler and far easier to defend, and much of the early criticism of Andreessen’s essay has accused him of falling into the same trap as the baptists and bootleggers he lambasts. By overstating the case and declaring that these things will happen, rather than may happen, many feel that he is adopting a position of faith.
Midjourney's rather psychedelic vision of a utopian AI future...
His views could also be seen as motivated by self-interest. As well as being a venture capitalist with investments in many businesses that stand to profit from unfettered AI development, Andreessen is on the board of directors of Meta, which is one of the leading developers of open source AI. This places him on one side of a long-standing divide in technology which has now been ported to the AI sphere, with the ironically named OpenAI forming the main opposition. Open source evangelists believe sunlight is the best disinfectant, and abhor the approach of OpenAI, who refuse to even discuss the size of GPT-4’s training data, much less what it contains. OpenAI claims this is to prevent a potentially harmful escalation in the AI arms race, but they also have the biggest gun, and their critics feel OpenAI’s pleas to be regulated are as much about solidifying their advantage as they are about protecting humanity’s future.
PREVENTION VS CURE
While it’s regrettable that such cynical narratives are deemed relevant, it’s harder to explain the more extreme positions on either end of the spectrum without them. Most people agree that AI could potentially develop into an existential threat to humanity, and that it could also help to cure disease, heal societies and avert climate disaster, but neither outcome seems probable. As Yann LeCun points out, this technology does not yet exist, and is unlikely to exist any time soon. The same is true of Andreessen’s utopian AI future; we do not currently have the means of arriving at it, and it is no less speculative than the doomsday scenario it seeks to counter.
...and, for balance, here's Midjourney's vision of "AI enslaving humanity".
So, will AI save us or enslave us? Is it boom or doom that we see coming down the tracks? There are no easy answers, of course, but perhaps we should be asking a different question: Is prevention really better than cure? The answer may depend on how much of the best possible outcome we’re willing to sacrifice, in order to prevent the worst possible outcome. Worryingly, there’s no real way to be sure, but getting the answer right could well be critical.
This story is part of our weekly briefing. Sign up to receive the FREE briefing to your inbox.
Gadgets & Gifts
Put your spy skills to work with these fabulous choices from secret notepads & invisible inks to Hacker hoodies & high-tech handbags. We also have an exceptional range of rare spy books, including many signed first editions.
We all have valuable spy skills - your mission is to discover yours. See if you have what it takes to be a secret agent, with our authentic spy skills evaluation* developed by a former Head of Training at British Intelligence. It's FREE so share & compare with friends now!
