“The AI on the horizon looks like Amazon Web Services — cheap, reliable, industrial-grade digital smartness running behind everything, and almost invisible except when it blinks off. The common utility will serve you as much IQ as you want but no more than you need. You’ll simply plug into the grid and get AI as if it was electricity. It will enliven inert object, much as electricity did more than a century past (..) There is almost nothing we can think of that cannot be made new, different, or more valuable by diffusing it with some extra IQ. In fact, the business plans of the next 10.000 startups are easy to forecast: Take X and add AI. Find something that can be made better by adding online smartness to it.”
- Kevin Kelly, The Inevitable (2016)
Understanding AI
I suppose a lot of people today mainly associate AI with generation of text, images, video, music, and the potential for massive job displacements. Stories about AI in these contexts, make their way to the headlines of newspapers and fill the top of social media feeds every single day.
However, the applications of AI are infinitely wider, deeper (and weirder) than generative AI and the threats to white-collar jobs. Perhaps there should be more focus on the occult and mind-expanding use cases of AI such as the quest to decipher ancient languages with AI, or to translate animal and plant communication. Another one of my favorites is NVIDA’s Earth-2 project, a digital twin of planet Earth that is being built to help us monitor and predict climate changes and natural disasters decades ahead of time.
With so many applications, it does not make sense to only talk about AI in the context of what it can do. We should have a common way of understanding what AI really IS in a way that is not context-specific and encompasses all the different things we can do with the technology.
Which leads us to the question: what IS artificial intelligence really? Especially the “intelligence” part is troubling and debated among linguistic experts. Inspired by Kevin Kelly’s quote above, I think of AI as “digital smartness on demand”.
Now, or very soon depending on how you see it, each and every one of us will have access to smartness on demand. It’ll be as impactful, ubiquitous, and freely accessible as electricity or the internet. In a few years’ time when the hype surrounding AI has died down, the technology will fade into the background scene of our lives and we will not even notice it. Such is the case with every other “general-purpose technology” (GPT for short, same acronyms as General Pre-trained Transformer — I bet OpenAI was aware of that when they promoted the term).
Understanding Smartness in the Education Sector
The education system uses performance in tests to evaluate how smart students are, or rather how smart they appear to be. In law school — from my own experience — certain cliques of ambitious law students tended to have an unhealthy fixation on grades since the big, prestigious law firms make a first screening of applicants, solely based on the merits of grades.
Now, we have GPT-4 that can ace any US college entry exam, including the uniform bar exam where GPT-4 performance corresponds to the top 10% of students. A liberal art student tested if ChatGPT-4 could pass her freshman year at Harvard, and it could do so with a median of A’s in its report card. GPT-4 is indeed very smart!
Smartness is probably a better term than intelligence here. After all, is it a sign of true intelligence for someone to pass a test or achieve a high grade? This “smartness” that we are used to evaluate students on can now figuratively be delivered in a bottle and we do not even have to pay for it.
Obviously, one way of dealing with students’ use of AI during their education would be to treat it as a problem and choose to examine students orally or “on premise”. Additionally, fraud detection tools like OpenAI’s classifier will become more powerful with every passing month.
However, the “death of the college essay” does not only present a problem for educators but also a huge opportunity. An opportunity to shift the focus away from getting good grades in school and towards building critical thinking and problem-solving skills that students can actually use to make a positive impact, for themselves and others, in the real world.
Paul Graham wrote an essay, The Lesson to Unlearn, about how succeeding in school essentially boils down to hacking bad tests. As a result, young apprentice startup founders tend to view business and life in the same way. In Graham’s experience, new graduates overcomplicate things and are often times surprised when he tells them that the way to make a good company is simply by making a good product. Instead, they think there is some sort of “trick” to it, like knowing the right people or launching a product on a certain day of the week. That is because, Graham infers, the education system has taught them that the way to win is by hacking bad tests. By finding the shortcut; the “hacks”.
In the best-case scenario, the future of education will be less formal and more substantial, less about chasing good grades and passing mandatory courses, and more about discovery, building character, and developing critical-thinking skills.
