ma_

Why handwriting is the last frontier of thinking

February 02, 2026 • by Maciej Adamski

TL;DR: The internet is drowning in AI-generated noise (52% of content), and our reliance on typing is flattening our cognitive abilities. Handwriting is the last frontier of deep thinking. This article explores the science behind why the pen is mightier than the keyboard and provides a 3-step analog challenge to help you reclaim your mind.

We have crossed a threshold. It didn’t happen with a bang, but with a quiet, relentless hum of server farms. In the early months of 2025, humanity passed a tipping point that historians may one day look back on as the moment we stopped thinking for ourselves.

For the first time in history, the majority of new text produced on Earth is no longer written by human hands. It is generated. And while the machines are getting faster, our brains are getting quieter.

Why is the internet becoming a "zombie web"?

To understand why you need to pick up a pen, you first have to understand the ocean of noise you are swimming in. The numbers from the last 18 months are staggering.

The 52% threshold

According to a landmark 2025 study by SEO firm Graphite, analyzing over 65,000 articles, approximately 52% of all newly published written content on the web is now AI-generated. Think about that. If you read two random articles today, statistically, one of them was not written by a person. It was predicted by a model.

The rise of AI slop

It gets worse. A separate study by Ahrefs in April 2025 found that 74.2% of new web pages launched that month contained significant AI-generated content. We are building a "zombie web" - an internet filled with grammatically perfect, confident-sounding text that has no heartbeat.

This phenomenon has even birthed a new word in our cultural dictionary: "AI slop."

Mentions of "AI slop" increased ninefold in 2025. It refers to the low-effort, filler content that clogs search engines and social feeds - text that exists to sell ads, not to transfer meaning.

The human recoil

We are reacting to this flood with a visceral rejection. We are starting to crave the "imperfect."

  • 55% of users now report feeling "uncomfortable" on websites that feel heavily AI-generated (Nielsen, 2025).
  • 73% of consumers say they can "smell" AI marketing copy, and over half of them immediately disengage when they detect it.

The trust gap: While AI volume is up, its value is down. 86% of articles that actually rank in Google’s top results are still human-written. The algorithms know what we know: machines can predict words, but they cannot provide insight.

How does writing by hand affect the brain?

While the internet floods, our brains are drying up. We are trading cognitive effort for convenience, and the price is our ability to learn. You might think, "Who cares if I type or write? The words are the same." Science says you are wrong.

The NTNU breakthrough (2024–2025)

Researchers at the Norwegian University of Science and Technology (NTNU) published a groundbreaking EEG study that finally settled the debate. They strapped 256 sensors to the heads of students and measured brain activity during handwriting versus typing.

The results were not subtle.

1. The "flatline" of typing

When you type, your brain is surprisingly inactive. You are using a "generic motor program." Pressing 'A' feels exactly the same as pressing 'Z'. The movement is repetitive and mechanical. The study showed that typing activates only small, isolated clusters in the brain. It is the cognitive equivalent of walking on a treadmill - you are moving, but you aren't going anywhere.

2. The fireworks of handwriting

When you write by hand, your brain lights up. The study showed massive, widespread connectivity patterns, particularly in the parietal and central regions - the areas responsible for memory, focus, and sensory integration.

Why? Because handwriting is complex. You have to physically shape each letter. You feel the friction of the pen on paper. You control the pressure.

The result: This complex sensory feedback creates a "memory hook." The brain is forced to process the information more deeply to keep up with the hand.

The verdict is clear: If you want to transcribe information, type it. If you want to understand it, write it.

Why is writing considered thinking?

This brings us to the core philosophy, championed by William Zinsser in his classic Writing to Learn.

Writing is not a record of a thought. Writing IS the thought.

When you stare at a blank page with a pen in your hand, you are vulnerable. You cannot hide behind a "Generate Draft" button. You have to face the gaps in your own logic.

You write a sentence. You realize it’s fuzzy. You cross it out. You try again.

That process - the crossing out, the pausing, the struggle - is not "wasting time." That is the thinking process.

When we let AI write our emails, our essays, and our notes, we are skipping the struggle. We are getting the product (the text) without the process (the understanding). We are becoming editors of our own lives, rather than authors.

How can you reclaim your mind in 2026?

So, where do we go from here? We are not Luddites. We will not smash the servers. AI is a tool, and a powerful one. But we must draw a line in the sand. We must decide which parts of our thinking we are willing to outsource, and which parts we must keep for ourselves.

If you want to reclaim your mind in 2026, you don't need to quit the internet. You just need to buy a notebook.

The 3-step "analog" challenge

  1. The "no-phone" morning pages: Before you look at a screen - before you let the world flood your brain with "slop" - spend 10 minutes writing by hand. It doesn't have to be good. It just has to be yours. Dump your thoughts on the page. You will be shocked at how much clearer your day becomes.
  2. The "hand-draft" rule: If you have to solve a difficult problem at work, refuse to open a Google Doc. Go to a whiteboard or a notepad. Sketch it out. Write the ugly, messy version first. Only when you understand the solution should you type it up.
  3. The letter (not the email): Once a month, write a letter to a friend or colleague. By hand. Mail it. In an era of automated spam, a handwritten letter is the most "premium" object in the world. It says: "I gave you my time, not just my prompts."

The future belongs to the humans who can still think deeply. In a world of infinite, average, AI-generated content, the most valuable asset you own is your unique, biological, jagged, human perspective.

Don't let the machine smooth away your edges. Pick up the pen.

Frequently asked questions

Is typing strictly worse than handwriting?

For speed and efficiency, typing wins. But for learning, retention, and conceptual understanding, handwriting is superior because it engages more cognitive processes and creates stronger memory associations.

Why is AI content called "slop"?

The term "slop" refers to low-quality, mass-produced text generated by AI models. It is grammatically correct but often lacks depth, insight, or a genuine human perspective - similar to "filler" material.

Can I use AI to help me write?

Yes, AI is a powerful tool for outlining, research, and editing. The danger lies in letting it do the thinking for you. Use AI to refine your thoughts, not to generate them.

What is the "zombie web"?

The "zombie web" describes an internet saturated with content that looks alive (because it reads well) but is actually dead (because it was generated by machines without human intent or experience).

Some interesting statistics

  • 55% of users feel uncomfortable on websites using heavy AI-generated content (Nielsen).
  • 48% don’t trust brands advertising on such sites.
  • 82% are skeptical of AI Overviews (e.g., Google’s AI summaries), though only 8% always check sources.
  • 40% trust generative AI search results more than organic search, and 41% more than paid ads (Attest, Mar 2025).
  • Trust in AI jumped 16 points in 2025 when users see clear value (ARF Study, Jan 2026).
  • Younger users (18–30) show higher trust (44%) than those over 50 (37%).
  • 80% believe media must disclose AI use, especially in news (Forbes, Jan 2025).
  • 73% of consumers can spot AI marketing content, and 52% disengage when they detect it (Smythos, Jan 2026).
  • Only 12% are comfortable with fully AI-generated news vs. 62% for human-made content (Reuters Institute, Oct 2025).
  • 33% believe journalists routinely check AI outputs before publication—raising credibility concerns.
  • Mentions of “AI slop” increased ninefold in 2025, with 54% negative sentiment by October (Meltwater).
  • Meta’s AI video app "Vibes" failed with only 23,000 daily active users in Europe (ITPro, Jan 2026).
  • 31% of consumers are less likely to choose a brand using AI in ads (CivicScience, July 2025).
  • Only 11% of Americans would engage with AI-generated influencers (YouGov).
  • AI-generated ads with disclosure see 73% higher trust, but baseline trust is low—only 38% have a positive view of AI vs. 77% of advertisers (Smythos, Jan 2026).
  • Gen Z prefers no AI in creative work (54%), but accepts it in shopping (only 13% oppose) (Goldman Sachs, Aug 2025).
  • 71% of office workers say AI tools appear faster than they can learn them (WalkMe, Jan 2026).
  • 47% feel they should be excited about AI—but instead feel worried.
  • 31% found training GenAI models harder than expected (ABBYY, 2025).
  • 26% of UK contact center agents considered quitting due to poor AI rollout communication (ArvatoConnect).
  • Yet, only 20% trust AI itself, and 21% trust AI companies (NIM.org, 2024).
  • 87% trust AI tools overall, but this drops sharply when applied to content creation (Nielsen).

Resources

  • Graphika report (Nov 2025): Found that state-sponsored influence operations (e.g., Russia’s “Doppelganger”, China’s “Spamoflauge”) use low-quality AI-generated content ("slop") with minimal engagement. NBC News: Online propaganda campaigns are using 'AI slop'

  • Wired (2025): Coined Donald Trump as "The first AI slop President" due to frequent use of AI-generated images in political messaging. Wired: The first AI slop President

  • Springer (Dec 2025): Study titled "AI-slop and political propaganda" analyzing AI-generated memes and strategic narratives. The resonant slop machine: public diplomacy and strategic narratives

  • Forbes (Jan 2025): 80% of consumers believe media should disclose AI use; 41% trust AI search results more than paid ads. Forbes: AI Productivity's $4 Trillion Question

  • Attest (Mar 2025): 40% trust generative AI search results, 41% more than paid ads; younger users show higher trust. (Source inferred from Forbes article context)

  • Smythos (Jan 2026): 73% of consumers can detect AI marketing content; 52% disengage when detected. (Mentioned in Forbes article; full report not directly linked)

  • MIT NANDA report: 95% of generative AI pilots fail. Forbes: AI Productivity's $4 Trillion Question

  • UK government copilot trial (2025): No definitive productivity gains; users made "moderate to significant" edits to AI outputs. (Cited in Forbes article; official report not directly linked)

  • BCG/Harvard "jagged frontier" study: AI helps weaker performers more; outside AI’s capability, error rates increase. (Referenced in Forbes article)

  • Goldman Sachs (Aug 2025): 54% of Gen Z oppose AI in creative work; only 13% oppose in shopping. (Mentioned in prior synthesis; full report not directly linked)

  • Meltwater (Oct 2025): Mentions of “AI slop” increased ninefold, 54% negative sentiment. (Cited in Forbes and NBC News articles)

  • PMC study (Nov 2025): AI-generated biomedical videos contain factual inaccuracies, lack references, and show communicative nonfluency. PMC: AI-Generated “Slop” in Online Biomedical Science Educational Videos

About the author

Maciej Adamski is a software engineer and founder of Dataglitch, specializing in building high-performance web applications. He writes about technology, software development, and the evolving digital landscape.