Given the growing backlash against AI, it’s not surprising to see a trope being invoked by some critics of AI. It goes something like this:
Trope: “We don’t need AI, we never asked for it, our lives are fine the way we were before AI….”
Perhaps a good number of people would agree with this trope, at least at a visceral level, especially given the recent concerns of a doomsday scenario involving human extinction and rogue AI agents over the past 3 weeks.
Yet, it’s important to unpack this trope to understand its logic.
new technologies pose a tradeoff
Every new disruptive technology poses tradeoffs, meaning it will likely involve potential benefits and harms. Part of the challenge is that, early on, humans cannot fully know or predict the extent of these benefits and harms, much less the secondary effects and innovation the technology will produce in the future. Faced with that lack of knowledge, it’s easy for one to jump to the conclusion that we “don’t need it.”
And, even after decades of a technology’s adoption by humans with much greater knowledge of those tradeoffs, there still will remain both benefits and harms resulting from the technology and posing stark choices. In hindsight, we might even be inclined to lament “we didn’t really need this technology, given the harms it produced,” but the ultimate decision on a technology’s adoption is ultimately left to everyone to make in market. If people don’t adopt a technology, it fails. If people adopt a technology, it succeeds.
humans don’t “need” electricity, either
Take electricity. Humans don’t really “need” electricity. There existed a more environmentally-friendly alternative: gas illumination and candles. In fact, if humans had not adopted electricity, arguably the world would be much better off in terms of CO2 emissions and less at risk of catastrophic climate change.
Using AI for assistance (even though I didn’t technically “need” to use it), I created this highly simplified comparison of the world we live in with electricity and a hypothetical world if humans had not adopted electricity and remained “the way things were” with gas-illumination. As Figure 1 below shows, in terms of CO2 emissions, the world would have been far better off without the adoption of electricity. That is, if CO2 emissions were the sole criterion. Indeed, we’d be much less close to catastrophic climate change under the Paris Agreement targets. This calculus doesn’t even consider all the additional societal ills electricity produces by powering, let’s say, addictive social media and, yes, AI. Maybe humans should have pulled the plug on electricity.
Figure 1. Comparison of CO2 Emissions in World With and Without Electricity
Of course, we have to consider some of the major societal benefits electricity yielded, including increased productivity and economic growth, advances in modern medicine, and, relatedly, increased human lifespans.
But that’s my point. How many people today are willing to give up electricity? Probably not many. And that is so even knowing that electricity greatly contributed to massive CO2 emissions by modern society that, according to some experts, has pushed the planet to the brink. Many of our concerns about extraordinary energy consumption would be solved, instantly, by the elimination of electricity. So would social media addiction and other ills made possible by electricity.
the stark tradeoffs with ai
The tradeoffs of AI might even be starker. The advances in curing cancer in the past few years (such as with Merck and Moderna’s treatment for skin cancer in a Phase 3 trial) have been promising. The Cancer Research Institute stated: “While the fight against cancer is an ongoing challenge, AI is helping to reshape how we understand, diagnose, and treat it in ways never before thought possible.”
AI is also advancing research in developing new drugs, as well as analyzing diseases, such as Parkinson’s, that have proven intractable to researchers. As Michele Vendruscolo, professor in biophysics and co-director of the Centre for Misfolding Diseases at the University of Cambridge, explained: “We can analyse this data and make very accurate predictions about the way candidate molecules will bind to the target at a scale that was unthinkable until a few years ago.”
Yet, at the same time, the same AI might enable bad actors to develop biological weapons. That’s a tradeoff.
other modern technologies pose tradeoffs, too
I’ve focused on electricity. But we could run this same tradeoff analysis for many common technologies that people have no hesitation in using today, including plastics, cars, trains, and airplanes. Few people in modern societies can say they don’t use any of these technologies.
Taking 1 cross-country flight consumes as much energy as 5 million AI text prompts (according to the analysis I ran using AI). So the average American can have a far greater impact on energy conservation by refusing to fly as opposed to refusing to use AI.
Taking 1 cross-country flight consumes as much energy as 5 million text prompts (according to the analysis I ran using AI).
So the average American can have a far greater impact on energy conservation by refusing to fly as opposed to refusing to use AI.
None of this is meant to diminish the potential harms and risks posed by AI. But it’s important to appreciate that every disruptive technology poses stark tradeoffs. And the “we don’t need this technology” trope can be said for every disruptive technology since at least electricity.
Electricity was a technology that humans never “needed.” Indeed, humans survived for generations with candles.
As the German art historian Wilhelm Hausenstein wrote during World War II, “Candles give as much light as things need in order to be what they are – optimally, so to speak – and allows them to retain their poetic element.” (Wilhelm Hausenstein, Licht unter dem Horizont, Tagebücher vxon 1942 bis 194 (Munich: Prestel Verlag, 1967), quoted in Schivelbusch, Disenchanted Night, p. 178.)
