The magic school of the future is one that helps students learn to work together and care for each other
AI is going to alter economic and professional structures. Will we alter the educational structures?
|Our Comprehensive Report | | Debate | Our Book
__
For the past 150 years, society has highly valued “intelligence,” which has been commonly defined as possessing a high IQ, extensive knowledge in one's field, the ability to identify patterns that others might miss, and strong analytical skills.
People who possessed this intelligence became “knowledge workers” and they were in high demand because, until recently, there were few methods to gain the benefits of such intelligence: hiring highly intelligent individuals, often at a high cost due to supply and demand; forming friendships with them; marrying someone possessing it; or, historically, through the unethical practice of enslavement (though enslavement was focused more on forced labor than forced intelligence).
In the context of employment, typically, the more intelligent a person is, the higher the salary they can command. This is evident in professions like CEOs, CFOs, brain surgeons, top academics, and lawyers, where high intelligence commands higher pay.
Now, machines can do many of the things we’ve traditionally defined as “intelligent,” and their ability to do so will grow rapidly. They already exceed human ability in these areas in many ways (no one knows as much as any AI across fields, and no one can discern patterns, analyze information, and make predictions as well as AIs can). If you need this type of intelligence, you can buy it, and you can buy it very, very cheaply.
Ross Anderson explains:
He (Sam Altman, CEO of OpenAI (ChatGPT)) has previously said that he expects the “marginal cost of intelligence” to fall very close to zero within 10 years. The earning power of many, many workers would be drastically reduced in that scenario.
Fabian Koh, writing in Channel News Asia, adds:
SINGAPORE: The rise of artificial intelligence (AI) will cause a “positive” shift in the traditional hierarchy of jobs, President Tharman Shanmugaratnam said on Wednesday (Nov 15). “What you regard as a better job and a job that deserves better pay might shift considerably in the years to come. The way we value IQ over EQ might change,” he said during a fireside chat on the first day of this year’s Singapore FinTech Festival (SFF). There is a range of jobs which require intelligence quotient, or IQ, that can actually be taken over by large language models (LLMs) and AI, and performed even more efficiently and better, said Mr. Tharman at the Singapore Expo.
What does this mean for the future of work?
It means some people are going to be in for some rough times. Why?
1. The unique value of having the highest intelligence and being the best will collapse. The reality is that you’ll be able to buy the level of “intelligence” you need for a fraction of the price of hiring most people and that intelligence will get you very far. If the intelligence you buy “augments” you, it basically allows you to “level—up” in a way that someone who is “more intelligent” than you cannot.
For example, a recent study of Boston Consulting Group employees found that those who used ChatGPT with minimal to no training radically improved their productivity as long as they used it for something that it was meant to be used for. The weakest employees gained the most, closing the gap with their more talented peers, who didn’t experience that much of a gain using it. AI works to level the unique value of intelligence in organizations.
2. The law of supply and demand related to intelligence will fall, and wages will decline. There is this huge, repetitive debate about whether AI will replace or create new jobs (or if it will create at least as many as it destroys). It’s easy to find thousands of articles that make strong (overconfidently sounding like ChatGPT :)) claims on one side of the debate, despite the fact that we’ve never had a piece of technology like this in human history and that it’s easy to find societies in the world that are so poorly structured (and AI will alter economic and political structures) that they have massive unemployment.
A lot of people will likely lose their jobs, at least in the transition to the AI World.
Basically, there are six themes in the knowledge worker unemployment debate.
(a) AI can do many jobs that people do, triggering unemployment.
There is basically a consensus that this will happen to a degree (Goldman Sachs says 7% of current jobs will be gone by 2030); the real debate is about the extent (some argue way more than 7%) and how many of those jobs will be replaced with new jobs, including those we haven’t even thought of.
People who strongly believe we are facing a significant wave of unemployment
Sam Altman (CEO, OpenAI)
Mira Murati (CTO, OpenAI)
Elon Musk (CEO x.ai; Tesla)
Dennis Hassabis (CEO & Co-Founder, Deep Mind (now part of Google)).
(b) If you use AI, you’ll be great at your job, and you won’t lose it
There is a consensus that most people will have to know how to use AI well to remain employed, but it’s far from certain that there will be jobs for everyone who knows how to use it well. How many bookkeepers using AI does a company need? And, most importantly, in my mind, what will your value be if AI can do most of your job? How many “less intelligent” and “less educated” people will be able to do your job if they can get massive assistance from an AI? How many more people will be able to be community doctors in a world where the AI assistant has all the needed medical knowledge for a basic, lower-stakes diagnosis? If there are millions of more people who can fulfill the role of community doctors (and eventually specialists), doctors won’t have much wage bargaining power (and neither will any of us).
© Massive economic gains will produce new economic activity and new jobs.
There is quite strong historical evidence that technology tends to improve productivity and that productivity improves economic growth. Higher levels of economic growth create more demand. More demand means people buy more things, and people think of new things to sell. Therefore, even if there is a disruption in employment, more demand and greater economic activity will make up for it. Bank of America predicts a $15 trillion dollar gain to the global economy from AI (by comparison, Germany’s GDP is around $4 trillion), and I’ve even seen a claim of $4 quadrillion if we hit AGI.
(d) In the short-term, this could hurt, but it will be okay [with proper action]
It’s sort of impossible to deny that at least the beginnings of new technological revolutions trigger massive unemployment and declining wages; simply look at the first wave of industrialization in ---. We will see jobs entirely automated or people becoming superfluous.
Even great optimists, such as Yann LeCun (AI “Godfather,” Chief Scientist @ Meta) who are convinced this will all work out, argue that we need social support structures and retraining.
Note: Google recently posted a blog arguing for the creation of a social adjustment fund for the unemployed and a reskilling program).
AI presents immense opportunities to catapult economies forward through increased productivity and economic activity that can benefit everyone. But we know that AI can also be a disruptive force. Industry, civil society, and policymakers will have to work together to prepare the workforce for the AI-powered job transition. One way they could do that is by creating a global AI Corps to build an AI-ready workforce by extending AI training programs to communities while helping workers impacted by AI get the skills needed to quickly bounce back to new and better jobs. Governments should also incorporate AI as a core component of education, create flexible immigration pathways for AI experts, and establish a safety net with an AI adjustment assessment program with tailored skilling programs. Senator Cantwell’s proposed “GI Bill for AI” in the U.S. and Singapore’s nationwide AI skilling initiative could serve as strong models for these types of global skilling programs.
This is something that our society will really struggle with. The current US Congress doesn’t want to spend another dollar, can barely keep the government open, and is trying to avoid fistfighting. Local legislatures and school boards are focused on what books to ban that students don’t read in the library. This is going to be rough.
I challenge you to find anyone who argues this will all just “work out.”
(e) What is really required is a significant re-organization of schooling and curriculum. At a meta-level, the school system is focused on developing the type of intelligence I opened with, and the economic value of that is going to rapidly decline.
(f). This is all going to happen very quickly (faster than any previous change in history), and many people aren’t paying attention. AI is already here.
Our Challenges
Traditionally, we’ve relied on the education system to help weather storms such as this, and this is going to be very difficult as we are undergoing a massive societal transition and it will be hard for education to adapt.
These are the challenges I see it facing.
1. Getting students to use AI tools. At least in K–12, to really facilitate widespread use, you have to ‘authorize” access to the tools, and many are not trying to facilitate access (certainly not all, as there are many public and private schools nationwide that are encouraging students to use these tools on their school networks and devices). Given that almost every job will require students to use AI tools, and they will certainly have to if they want to build their own businesses, the education system needs to figure out how to get students to use these tools and support them in doing so. It’s hard to think of more than a handful of things that are more important to teach students about. One pro tip: just find a way to let your students use “New Bing” that has AI integrated. It’s free, and it’s easy to use.
2. Training. Faculty and staff need to be trained on these technologies. There are online options, but many need human support, and many, despite some popular thinking in this area, cannot figure it out on their own.
3. Resistance. Some people don’t like this stuff (and some haven’t even tried it) so they will need to be coached, encouraged, incentivized, etc.
4. AI literacy. People need to understand how much and how quickly the world is changing/is about to change.
5. Reorient the education system. Although schools are struggling with how to manage them now, 1-3 are really the easy ones to pull off and they present the best prospects for people to be able to use the tools to amplify their intelligence.
Four is essential to live in this world.
.The more difficult challenge is that we need to move the education system away from such a great emphasis on teaching content and analysis of content, as that type of “intelligence” isn’t going to have much economic value anymore. And, as Harvard Professor Chris Dede notes, moving to use AI in schools to simply teach students how to do what machines can already do better than us is a “nightmare,” and it’s a nightmare that’s starting to become reality.
We need to focus on helping students develop skills around communication, collaboration, critical thinking, and creativity (these will all be needed in humans even if the AI eventually develops them) and develop these skills (and relevant content knowledge) through project-based, interdisciplinary, and immersive/experiential learning experiences so they can fully develop and thrive as individuals who will co-exist with people (and robots) in a transformed world.
And we need to work with them to develop skills in the “human” sciences, helping them learn to care for others (friends, parents, and young children).
Stuart Russell notes that “most human beings are going to be in these interpersonal roles” where people care for one another, including caring for children, parents, and friends.
And we want to help children have courage, to have convictions, and to persevere. We humans will need these even if machines develop them. Machines can’t replace our courage, convictions, perseverance, and love.
These are things that we humans can do (though I need to be much better at it).
What do we need to do?
AI is going to alter the economic structure.
AI is going to alter the “professions.”
We need to alter the education system if it’s going to remain relevant, and we need to start preparing students for this world immediately. It’s already here. Where are we? Rearranging the chairs?
Do schools have to do this?
Do schools have to focus on 1-5? No, they don’t. But if they don’t, they shouldn’t complain if kids don’t show up and are disengaged after they learn the basics of reading, writing, arithmetic, and the basics of how their society and economy work. They will still need to learn some “information,” how to organize it, etc., but as a focus of education, it’s not going to make a lot of sense, as it will have very little economic value in the near future.
The magic school of the future is one that helps students learn to work together and care for each other, and it’s something that people who work in schools are prepared to do.
__
Some jobs will be created, but just like every major revolution, I think a lot of jobs will be lost. There probably be a bigger impact on jobs than in any other revolution, and we have to prepare for this new way of life...Maybe we work much less. Maybe the workweek changes entirely.
Mira Murati, CTO, Open AI
(J)obs are definitely going to go away, full-stop.
Sam Altman, CEO, OpenAI
You can have a job if you want.
Elon Musk, CEO X.ai