RESEARCH / SOURCE GUIDE

Read the evidence.
Keep its limits.

Selected original studies and institutional sources behind the book’s central questions, with the boundaries that matter.

What a source can say
is part of the story.

A paper about insects is not a paper about humans. A study linking screen use with well-being is not automatically a causal test. A short intervention is not a lifetime forecast. This guide points to primary research and responsible institutions so you can inspect the claims for yourself.

The book contains 28 numbered source entries. This page highlights twelve that anchor its main argument; it does not certify every sentence or replace a systematic review.

SELECTED SOURCE NOTES

Use the original link, then ask what was measured, in whom, for how long, and what alternative explanations remain.

  1. 01

    Sea turtles & artificial light

    Florida Fish and Wildlife Conservation Commission

    Hatchlings can be disoriented by artificial lighting. This supports the beach cue example, not a shared human mechanism.

    Read the original source (opens a new tab)
  2. 02

    Darwin’s plant experiments

    Charles and Francis Darwin, 1880

    Original observations on seedlings and light sensitivity. Plant movement does not establish human desire.

    Read the original source (opens a new tab)
  3. 03

    Why insects gather near light

    Fabian and colleagues, 2024

    Flight tracking supports a dorsal-orientation account for the insects and conditions studied.

    Read the original source (opens a new tab)
  4. 04

    Natural selection

    Natural History Museum

    A clear introduction to inherited variation and selection, plus the history around Darwin.

    Read the original source (opens a new tab)
  5. 05

    Evolutionary rescue

    Graham Bell, 2017

    Adaptation can sometimes aid persistence. It does not guarantee rescue from changing environments.

    Read the original source (opens a new tab)
  6. 06

    Culture evolves too

    Creanza, Kolodny & Feldman, 2017

    Cultural transmission changes what people learn and pass on. It does not guarantee progress.

    Read the original source (opens a new tab)
  7. 07

    Reward prediction

    Schultz, Dayan & Montague, 1997

    Foundational reward-prediction work. It is not a direct experiment on modern phone use.

    Read the original source (opens a new tab)
  8. 08

    Evening eReaders & sleep

    Chang and colleagues, 2015

    Controlled evening exposure shows specific sleep and circadian effects; it does not indict every digital activity.

    Read the original source (opens a new tab)
  9. 09

    Adolescent well-being

    Orben & Przybylski, 2019

    Large dataset analysis of associations. Association alone does not establish causation.

    Read the original source (opens a new tab)
  10. 10

    Facebook deactivation

    Allcott and colleagues, 2020

    A four-week experiment found well-being gains alongside costs to news knowledge.

    Read the original source (opens a new tab)
  11. 11

    Blocking mobile internet

    Castelo and colleagues, 2025

    A short intervention reports benefits with important adherence, selection, and follow-up limits.

    Read the original source (opens a new tab)
  12. 12

    Social media & adolescent health

    National Academies, 2023

    Benefits and harms warrant attention; population-level causal claims remain difficult.

    Read the original source (opens a new tab)
What is an interpretation?

The site’s central distinction—attention is a visible signal; benefit must be examined—is an interpretation built from the book’s question and sources. It is not a single finding established by one experiment.

Approved cover of Almost Completely Written By AI
A BOOK BEHIND THE QUESTIONS

Follow the idea
into the book.

Almost Completely Written By AI connects sea turtles, Darwin, attention, and the choices people can make now. The pages shown here are abstract artwork; the manuscript is not reproduced.

See the cover, facts, and editions →