All articles

Is AI bad? An honest answer from people who build it

Andrew Powers
Andrew Powers··5 min read
A bear holding a lantern stands where a path splits toward a meadow and a thicket.

AI isn’t bad by itself, but most worries about it are fair. The problems come from how it was built and who owns it. It learned from other people’s work, squeezes entry-level jobs, makes work look the same, uses a lot of power, and pays most gains to a few companies.

Most people worry, and with reason

Half of Americans say AI makes them more concerned than excited, Pew found, and only one in ten say the opposite. We build AI tools at PageLines, and we find most of those worries reasonable.

AI learned from people’s work without asking

AI models learned from books, articles, music, and art, mostly without the creators’ permission. Some of that material was pirated, and Anthropic agreed last year to pay authors $1.5 billion over pirated books. Courts have mostly accepted the companies’ fair-use argument, but the creators still lost their work. Why AI companies will keep training on data they don’t own explains why this will continue.

AI squeezes entry-level jobs

The job losses fall unevenly. Stanford researchers found that since late 2022, employers hire fewer young workers in the jobs AI does best, while experienced workers in those jobs keep growing. AI does the tasks beginners used to learn on, so the first job that leads to the second gets harder to find.

AI makes everyone’s work look the same

When everyone uses the same AI, their work starts to look alike. Studies of brainstorming, consulting, and writing found one pattern: AI improves each person’s work and makes the group’s work more similar. That sameness drives much of the fatigue with “AI slop.” Asymmetric AI covers what to do about it.

AI uses a lot of electricity

AI runs in data centers, and data centers used about 1.5% of the world’s electricity in 2024. The International Energy Agency expects that to more than double by 2030, mostly because of AI. The cost is real, though AI still drives a small share of the world’s growth in electricity use.

A few companies collect most of the gains

Control may be the biggest worry. Amazon, Microsoft, Alphabet, and Meta plan to spend more than $700 billion on AI this year, and most people rent the same few models from them. That concentration reaches government surveillance, where company policies set the main limits on AI.

Some fears outrun the evidence

So far, job losses hit early-career roles, not the whole economy. And AI has no goals of its own. It makes confident mistakes, so check its work rather than fear it.

AI should serve the person using it

Opposing AI won’t stop anyone from building it, so ask who it serves. We think AI should make people more capable and stay under their control, not replace them or report to someone else. So we build on open models, give each person an AI that works only for them, and put a community of people ahead of the software.

Bottom line: AI isn’t bad by itself, but the worries about copying, jobs, sameness, energy, and power are mostly fair, and most come down to who controls it. Use AI that works for you, keep your own judgment, and stay close to people who tell you the truth. PageLines is a community of AI leaders with an AI of your own on open models. See how it works.

Common questions

Is AI bad for the environment?

AI uses a lot of electricity. Data centers used about 1.5% of the world's electricity in 2024, and the International Energy Agency expects that to more than double by 2030, mostly because of AI. The cost is real, though AI still drives a small share of growth in electricity demand.

Is AI taking jobs?

So far the pressure falls mostly on entry-level work. Stanford researchers found that young workers in the jobs AI does best have been hired less since late 2022, while experienced workers in the same jobs have kept growing.

Why do people hate AI?

Most dislike comes from fair concerns: AI was trained on people's work without permission, it is squeezing entry-level jobs, it floods the internet with low-quality content, and its gains go to a few large companies.

Is it wrong to use AI?

No. Ask who controls it, what it learns from you, and whether it makes you more capable or more dependent. AI you control, plus your own judgment, answers most of those.