Every technology has downsides. Paper, gunpowder, and the printing press advanced civilization, but also meant societal instability and conflict for humanity. More recently, we’ve experienced for ourselves the double-edged swords that are the internet, mobile computing, and social media.

What are the downsides that come with the use of AI, and how do we manage them?

Telling Us What to Think

Is our tech telling us what to think?

Our thoughts are influenced when we endlessly scroll social feeds curated by attention-retaining algorithms, when we see content from web searches tuned and optimized to ensure they reach us, and now also when we experience interacting with LLMs. The internet has enabled an explosion of information, along with unprecedented access, while allowing anyone to say anything.

It’s important not to confuse acquiring new information with thinking. Being told an answer—if it’s the right one at all—is different from the independent ability to arrive at the answer.

Using tech influences the way we think, how do things, and what we create.

Telling Us How to Nurture

AI is breaking the bottom rung of the career ladder for young workers in some industries.

By replacing routine work, it’s also replacing the need for many internships and entry-level roles. Employers increasingly prefer to hire experienced workers over fresh graduates. The learning that usually happens in these roles then doesn’t get to happen, disrupting how we traditionally help junior employees gain experience.

AI is affecting how we pass on our knowledge and experience—training, mentoring, and investing in the next generation.

Telling Us What to Do

Sit down in front of a blank sheet of paper. What does it invite?

A keyboard invites us to type. A touchscreen invites us to swipe. A device notification invites our attention.

AI invites use of it for everything cognitive.

This means that the downsides that come along with the numerous uses for AI are too many to list. So far, we’ve seen: security risks, personal data risks, software bugs being easier to both fix and exploit, health risks, overinvestment and financial risks, valuation bubbles and economic risks, regulatory and policy risks, environmental downsides, and risks to society.

Use of technology comes with immediate side effects, systemic dependencies, and difficult to predict downstream ripple effects. On top of that, our environment, systems, and the technology itself will keep evolving. The downsides to be managed will continue changing too.

How then, should we try to manage these downsides?

See Tech For What It Is

I think the solution is twofold.

First, we need to comprehend the real value of AI. What does it allow us to do, and what does it do for us? It should be viewed in its entirety—both the pros and cons. It should be seen for its actual capabilities, not the projected or hypothetical ones.

Whenever new technology arrives, our defaults surreptitiously shift to the new “standard”. We’ve seen the “Find us on Facebook / Talk to us on Twitter” stickers on every company truck. And who even uses paper anymore?

Holding a hammer, everything looks like a nail. Technology is a double-edged sword. We must know what is gained wielding a new technology, and what is lost. We have to be cognizant of the negatives so we may minimize them.

Adopting AI should be a decision we make with open eyes, not one to sleepwalk into. We have to be conscious of: What are the upsides, and is the “upgrade” worth the tradeoff for a particular use? Choices should be made purposefully, not mindlessly. They should be more considered, especially within the many companies making their employees adopt AI.

Trouble brews for us when this decision isn’t made consciously, or gets made by others on our behalf—no matter if due to a systemic cause, under influence or pressure, or because of some external incentive.

Know What We Want

Secondly, and ultimately, we need to know what we want.

Take these forms of communication, which use varying degrees of technology: in-person conversation, FaceTime, phone call, snail mail, email, Slack, text, and AI chatbot. Each has its own time and place to be used, depending on factors like: who’s being spoken to, what’s being said, or what kind of relationship is it. We choose the right tool for the task, in view of the result we want.

AI lets us do many things. It makes things easier to both do and delegate. It lets us do more faster, by doing less. We could spend that freed up attention elsewhere, or back on the same task. We could work at a higher level of detail and complexity than before. We could take on more and be more ambitious, if we wanted—but what do we want?

Going faster only matters if we’re heading in the right direction. How a technology leans may not necessarily be in sync with where we'd like to go. Swept up in a tech revolution, we find ourselves indiscriminately adopting it everywhere. We overengineer solutions to problems that don’t require them. When we don’t know what we want, we do busywork, and needlessly subject ourselves to the downsides.

It’s up to us to know and decide how to use technology. We have to know what we want.

Don’t Forget We’re Human

See AI in its entirety, and know what we want. Both easier said than done.

The negatives of AI require active mitigation, yet we’re overwhelmed and outgunned by them. Also, how many of us deliberately pursue experiences in exploration of personal purpose, let alone have one?

Layer on how the downsides from AI can feel like a marginalization of our humanity. It’s trained on works made by us—from our blood, sweat, and tears—to be able to produce its outputs. It’s replacing jobs, doing the work cheaply, effortlessly, and near-instantly.

So instead here’s a practical, perhaps radical, view: We should prioritize being human.

Again, not to eschew technology—which is impossible in modern life. But I think we can say it’s unhealthy to let technology entirely lead and decide your life for you.

Spend some time without tech. Even if only on occasion, we should try to make time for doing things in older, slower, harder ways, using lesser degrees of technology:

  • Read the whole article, instead of going straight to the comments.
  • Watch the whole interview, instead of only getting the AI summary.
  • Read books instead of watching Netflix.
  • Appreciate the beauty of nature in place of digital entertainment.
  • Do nothing and let your mind wander instead of scrolling on your phone.
  • Build relationships in-person instead of using social media.
  • Ride the bicycle around the park instead of running on a treadmill.
  • Write using your brain instead of using AI.
  • Go to sleep instead of “rest” by revenge-watching short-form content.
  • Sketch your thoughts out on paper, instead of in a notes app.
  • Maybe even start your own fire instead of cooking on the induction.

From time to time, decide to slow down. Engage in the fuller, higher fidelity life experience, instead of letting technology package it for you. These are opportunities to remind ourselves what technology can’t do, the limits of AI, and what we miss out on. To stave off the negative effects of tech, spend more time living a human life.

Many will, in response to this, say they can't afford the time to.

But really, I think we can't afford not to.

I write about doing my life's work, living up to my potential as a productivity solopreneur, and what AI means for my work: subscribe to THPC on Substack. This article was written with WriteMapper, a mind-mapping desktop app for thinking and writing, available on macOS and Windows.