Is AI dominance inevitable? A technology ethicist says no, actually

Anyone following rhetoric around artificial intelligence in recent years has heard one version or ar of claim that AI is inevitable. Common mes are that AI is alrey here, it is indispensable, and people who are bearish on it harm mselves.

In business world, AI vocates tell companies and workers that y will fall behind if y fail to integrate generative AI into ir operations. In sciences, AI vocates promise that AI will aid in curing hirto intractable diseases.

In higher education, AI promoters monish teachers that students must learn how to use AI or risk becoming uncompetitive when time comes to find a job.

And, in national security, AI’s champions say that eir nation invests heavily in AI weaponry, or it will be at a disvant vis-à-vis Chinese and Russians, who are alrey doing so.

argument across se different domains is essentially same: time for AI skepticism has come and gone. techlogy will shape future, wher you like it or t. You have choice to learn how to use it or be left out of that future. Anyone trying to stand in techlogy’s way is as hopeless as manual weavers who resisted mechanical looms in early 19th century.

In past few years, team at UMass Boston’s Applied Ethics Center have been studying ethical questions raised by widespre option of AI, and ethicist believes inevitability argument is misleing.

History and hindsight

In fact, this claim is most recent version of a deterministic view of techlogical development. It’s belief that invations are unstoppable once people start working on m. In or words, some genies don’t go back in ir bottles. best you can do is harness m to your good purposes.

This deterministic approach to tech has a long history. It’s been applied to influence of printing press, as well as to rise of automobiles and infrastructure y require, among or developments.

But ethicist believes that when it comes to AI, techlogical determinism argument is both exaggerated and oversimplified.

AI in field(s)

Consider contention that businesses can’t afford to stay out of AI game. In fact, case has yet to be me that AI is delivering significant productivity gains to firms that use it. A report in Ecomist in July 2024 suggests that so far, techlogy has h almost ecomic impact.

AI’s role in higher education is also still very much an open question. Though universities have, in past two years, invested heavily in AI-related initiatives, evidence suggests y may have jumped gun.

techlogy can serve as an interesting pedagogical tool. For example, creating a Plato chatbot that lets students have a text conversation with a bot posing as Plato is a cool gimmick.

But AI is alrey starting to displace some of best tools teachers have for assessment and for developing critical thinking, such as writing assignments. college essay is going way of disaurs as more teachers give up on ability to tell wher ir students are writing ir papers mselves. What’s cost-benefit argument for giving up on writing, an important and useful tritional skill? In sciences and in medicine, use of AI seems promising. Its role in understanding structure of proteins, for example, will likely be significant for curing diseases. techlogy is also transforming medical imaging and has been helpful in accelerating drug discovery process.

But excitement can become exaggerated. AI-based predictions about which cases of COVID-19 would become severe have roundly failed, and doctors rely excessively on techlogy’s diagstic ability, often against ir own better clinical judgment. And so, even in this area, where potential is great, AI’s ultimate impact is unclear.

In national security, argument for investing in AI development is compelling. Since stakes can be high, argument that if Chinese and Russians are developing AI-driven automous weapons, United States can’t afford to fall behind, has real purchase.

But a complete surrender to this form of reasoning, though tempting, is likely to le US to overlook disproportionate impact of se systems on nations that are too poor to participate in AI arms race. major powers could deploy techlogy in conflicts in se nations. And, just as significantly, this argument de-emphasises possibility of collaborating with versaries on limiting military AI systems, favouring arms race over arms control.

One step at a time

Surveying potential significance and risks of AI in se different domains merits some skepticism about techlogy. I believe that AI should be opted piecemeal and with a nuanced approach rar than subject to sweeping claims of inevitability. In developing this careful take, re are two things to keep in mind:

-First, companies and entrepreneurs working on artificial intelligence have an obvious interest in techlogy being perceived as inevitable and necessary, since y make a living from its option. It’s important to pay attention to who is making claims of inevitability, and why.

-Second, it’s worth taking a lesson from recent history. Over past 15 years, smartphones and social media apps that run on m came to be seen as a fact of life – a techlogy as transformative as it is inevitable. n data started emerging about mental health harms y cause teens, especially young girls. School districts across United States started to ban phones to protect attention spans and mental health of ir students. And some people have reverted to using flip phones as a quality of life change to avoid smartphones.

After a long experiment with mental health of kids, facilitated by claims of techlogical determinism, Americans changed course. What seemed fixed turned out to be alterable. re is still time to avoid repeating same mistake with artificial intelligence, which potentially could have larger consequences for society.