Is AI Making Us All Dumber?

We know artificial intelligence is here to stay.

From social media algorithms to customer service chatbots and machine-learning-based science, it seems that some form of AI is behind everything we see, hear, and do.

And this is obviously impacting the job market.

According to the latest Future of Jobs survey by the World Economic Forum, 86% of employers believe AI and information processing technologies will transform their business by 2030.

It’s not surprising, then, that employers expect skills in AI and Big Data to grow in importance over the coming years —  more than any other workforce skill.

The future of work is also the future of working with — and, eventually, for — AI. But what does that look like? And what happens to humans in the process?

Ironically, as AI becomes smarter every day, we might be getting dumber in the process.

Letting machines do the thinking for us

A recent study by Microsoft sounded the alarms. It would seem that, as knowledge workers place greater trust in artificial intelligence, they’re also exercising less critical thinking.

Another study, published in Societies, came to a similar conclusion. After surveying over 600 participants, researchers found a negative correlation between AI usage and critical thinking — especially among younger and less educated respondents.

The culprit? Cognitive offloading.

More and more, we’re delegating cognitive tasks to external tools such as GenAI. In theory, this can be a good thing. We can take a load off our minds and refocus our attention elsewhere.

But there’s a dark side to so much convenience. When users rely on AI to make decisions for them, they tend to make less of an effort to actually think things through on their own.

As one participant remarked: “It’s great to have all this information at my fingertips, but I sometimes worry that I’m not really learning or retaining anything. I rely so much on AI that I don’t think I’d know how to solve certain problems without it.”

When AI helps us do our jobs, we run the risk of forgetting or never learning how to do those jobs in the first place: a phenomenon called deskilling.

How we ended up here

We’ve heard this story before.

Computers have been making life easier for us for decades — while also making us less competent, aware, and knowledgeable.

Many years ago, a landmark study found that search engines and the internet are reshaping human memory. We’re recalling less information, because we know it’s online if we ever need it. Researchers called this the “Google effect.”

And it’s even affecting how we interpret the world around us.

Why bother learning how to get around your city, neighborhood, or community, when you can just check Google Maps? In fact, frequent use of GPS technology has been linked with a decline in spatial awareness.

Our overdependence on computers is a serious issue — and it can have serious consequences. Remember Air France 447?

Back in 2009, a routine flight from Rio de Janeiro to Paris encountered what should have been a minor complication. Due to weather conditions and instrument failure, the plane’s autopilot disengaged. The aircraft continued cruising at the right speed, in the right direction, at the right altitude: Nothing should have happened. But the human pilots panicked. Faced with misleading information on their flight displays, they overcorrected and descended into tragedy.

As aviation expert William Langewiesche explained, in his article for Vanity Fair:

“Automation has made it more and more unlikely that ordinary airline pilots will ever have to face a raw crisis in flight — but also more and more unlikely that they will be able to cope with such a crisis if one arises.”

This is how it usually goes with technological leaps: They push us forward, but they also affect our lives, our minds, and our society in ways we neither expect nor want.

Automobiles didn’t simply get us places faster. They also altered our perception of time and space, redrew the layout of our cities, and shifted the logistics of food distribution. They disengaged people from local economies and generated health and environmental problems that we’re still figuring out how to deal with.

Smartphones were similarly transformative. They impacted how we talk to one another — or don’t, for that matter. We are now always online, always available, and always reachable, so we don’t have to meet up in person. We can look up any information at any time and anywhere, so we don’t have to remember it or know it.

Such unforeseen consequences — or externalities — are perhaps an inevitable part of widely adopting innovative, life-changing technologies.

But that doesn’t mean we have to simply accept these consequences. On the contrary, we should take action so we can enjoy the benefits of our inventions while curtailing the negatives.

When applied thoughtfully, technology can improve our quality of life, making everything from shopping to using public restrooms easier and more comforting, without taking away our agency as users, citizens, or workers.

How we get back control

So, what’s the solution? After all, computers and AI aren’t going anywhere. They’re too pervasive, too useful, and — in today’s fast-paced world — too necessary.

In a 2023 survey of employees and business leaders, 74% of respondents admitted they had to make 10 times more decisions than a few years prior, while 70% of business leaders wished a robot could just make all those decisions for them.

Of course, AI probably shouldn’t be running our businesses. But it can still help us make better decisions.

As that aforementioned Microsoft study points out, AI tools have plenty to offer. They can organize information, offer personalized solutions, automate routine tasks, and much more. Thing is, in order to take advantage of these benefits, we also need human stewardship. We need people who know how to fine-tune prompts, steer AI responses, and verify facts or claims.

In short, we need to learn how, where, and when to use AI. Because when we land on that sweet spot between human oversight and AI assistance, great things can happen.

For example, experienced programmers have been found to be more productive when using CoPilot. Even when CoPilot makes mistakes — and it makes plenty — programmers can quickly catch and correct them. In fact, they might find it easier to fix CoPilot errors than human errors. The process ends up being faster than having humans write code from scratch.

But here’s the thing: Expertise takes many years to develop. To supervise AI outputs, there must be humans who know what they’re looking at. For today’s newbies to become tomorrow’s experts, they will need spaces where they can practice, fail, and learn without AI looking over their shoulders.

This idea is gaining traction. For example, the National Health Service in England recommends that, even as AI is integrated into healthcare workflows, a proportion of cases should still be handled manually for clinicians to develop and retain their skills and professional judgement.

After all, let’s not forget that, for the time being, we can’t entirely trust AI to call the shots. Or even read the news.

According to BBC research, if you ask AI about what’s going on in the world, there’s a 51% chance you’ll receive a misleading answer riddled with tiny issues, inconsistencies, misrepresentations, or major factual errors.

Even when it’s not lying to us, AI is easy to sway, influence, or break.

Spray graffiti on nearby stop signs? Autonomous vehicles may not be able to identify them anymore. Use emotionally-charged language when discussing a traumatic event? Your AI therapist may actually get stressed out.

Yet the way is forward, not backward. We shouldn’t fight against AI, but walk together.

Just as you can stress out your AI therapist — you can also calm them back down. The same study cited above also found that prompting mindfulness exercises or relaxation language helped regulate ChatGPT’s anxiety levels. So you see: Humans and AI can assist each other.

AI is part of our future, and we need to figure out where the path forward is heading. Because it won’t be our AI tools leading the way: It’ll be us, creating, taking risks, investing, suggesting solutions, and finding new problems.

In other words, we’ll still need to do most of the thinking. It’s our vision, our hopes, and our dreams that matter. AI can only help make them a reality.