Direct answers, on the record.

No. The foreword opens with the sentence "I wrote parts of this book using the exact machines I am warning you about." The book argues that AI is a pharmakon — a tool that is medicine or poison depending on the dose — and the 30-day protocol is the dosage guide. The author builds with AI every day and has trained more than 1,500 professionals and AI users. The argument is that you can use AI harder than most people without losing the cognitive tissue AI consumes in most people.

Yes — and the term is used deliberately. An AI detox in The Anti-AI Brain is a 30-day, neuroscience-first reloading of the four cognitive circuits that AI delegation thins out — attention, memory, reasoning, and decisions. It is not abstinence: you keep using AI throughout the protocol. It is not recovery from a clinical diagnosis: the book deliberately avoids the DSM-5-adjacent framing used by instruments like the Artificial Intelligence Addiction Scale (AIAS, 2025). It is a dosage protocol — closer in spirit to an athlete's off-season than to an addiction recovery program. The reader finishes the thirty days using AI harder than most people, while preserving the cognitive tissue that most people have already handed over.

Knowledge workers, researchers, writers, founders, engineers, students, and parents who use AI daily or daily-adjacent and have noticed either (a) a quiet decline in their own unassisted focus, memory, or decision-making, or (b) the same decline in someone they care about. The book is calibrated to the reader who cannot and will not stop using AI — quitting is not on the menu.

Newport's 2019 book is the category reference for intentional technology use and it sits in the lineage this book extends. Digital Minimalism addresses attention economics and screen-time discipline in a pre-LLM world — social media, notifications, task-switching. The Anti-AI Brain addresses the specific post-ChatGPT problem: what happens neurologically when the machine can produce the output for you, and how to preserve the cognitive circuits that delegation shrinks. Newport's answer is subtraction — a 30-day digital declutter, then re-introduce only what serves your values. This book's answer is dosed use — a 30-day neuroscience protocol that reloads specific circuits (attention, memory, reasoning, decisions) while you continue to use AI. Different mechanism, different research base (LLM cognitive-debt studies did not exist in 2019), different protocol. The books are complementary. Newport's work on attention and values underwrites the Lens layer of this book's five-layer model.

Those books were written before large language models. Deep Work (Newport, 2016) and Stolen Focus (Hari, 2022) focus on attention economics — the cost of notification-driven task-switching. The Shallows (Carr, 2010) addresses hyperlink-era reading. The Anti-AI Brain addresses the specific post-ChatGPT problem: what happens neurologically when a tool that can produce the output for you becomes ambient, and how to preserve the underlying circuits while continuing to use it. The mechanism is different, the research base is different, and the protocol is different.

Approximately 68,000 words — about 280 pages in paperback. Three parts: Part I names the condition (neuroscience synthesis), Part II is the 30-day protocol, Part III is the long game. Every chapter ends with a one-page Operator Brief that distills the action.

Yes. The flagship study is Kosmyna et al., 2025 at MIT Media Lab: an EEG study of 54 adults across repeated essay-writing sessions showed roughly 55% lower alpha-theta coupling in the prefrontal cortex of the ChatGPT group compared with the unassisted group, and more than 83% of ChatGPT users could not accurately recall a sentence from an essay they had just "written." Corroborating findings come from Microsoft Research (Lee & Sarkar, 2025, n=319 knowledge workers at CHI 2025), Wharton (Shaw & Nave, 2025), and a four-university RCT (CMU/Oxford/MIT/UCLA, 2026, n=1,222). The full evidence ledger is on the research page.

It is a preprint (arXiv, June 2025) currently undergoing peer review. The book flags this explicitly and does not rest any structural claim on that single study. The argument is built from roughly a dozen anchor studies across four independent labs — most of them published in Nature, Science, PNAS, or Psychological Review — so the thesis does not fall if the Kosmyna preprint changes on peer review.

Partially, and measurably. The adult brain is neuroplastic — Maguire’s 2000 PNAS paper on London taxi drivers demonstrated hippocampal structural change from navigation training in working adults, and the 2025 McGill "online brain-training reverses ten years of aging" result (ScienceDaily, October 2025) is one of several recent demonstrations that cognitive load in adulthood restructures tissue on measurable indices. Thirty days is enough to reinstall the behaviors and begin the tissue-level changes; the book’s Part III covers the twelve-month arc.

Not permanently — but the longer the circuits go unloaded, the more cognitively expensive the reload becomes. The research base for this is strong. Sparrow et al. (2011, Science) on the Google Effect, Ward et al. (2017) on smartphone-presence working-memory cost, and Madore et al. (2020, Nature) on media-multitasking memory failures all show that the relevant circuits are use-dependent. The book argues the urgency is not existential, it is compound: each untrained month raises the price of the next month’s training.

About 90 minutes. You already spend that much time on your phone before noon. The protocol reallocates the time rather than adding to your day. By Week 2, most readers report a net reduction in total screen time because the underlying work gets faster.

Yes. Chapter 11 ("The Outsourced Mind") is specifically for this case. The first two weeks feel distinctly unpleasant — the book names this the Ten-Minute Wall — and Part II explains why that feeling is the signal the protocol is working rather than a signal that you cannot do it. The reader tier most transformed by the protocol is the heavy-AI-use tier, not the light one.

Yes, with constraints. Week 1 is unassisted by design so you can see your actual baseline. Starting Chapter 7 (Day 9), the book installs the Socratic Mode Custom Instructions — a prompt layer that converts the model into a sparring partner instead of a ghostwriter — and from Day 20 onward you use AI on a structured schedule. By the end of the month you are expected to use AI better than before the protocol, not less.

The Anti-AI Brain Score quiz (free, on-site) places you in one of four reader tiers — Ghost, Partisan, Operator, Sovereign — and emails you a tier-specific Operator Brief. The book itself includes the full Anti-AI Seven worksheet, the 30-day checklist, and the Socratic Mode prompt pack. A standalone workbook and audio course are scheduled for Q3/Q4 2026.

Nik McFly is a writer and AI practitioner in Almaty, Kazakhstan. He has been shipping with machine-learning systems since 2020 and has trained more than 1,500 professionals and AI users. His other books — Vibe Patenting, The Polymarket Files, The Complete AIO Playbook, Digital Dolls, Real Dollars, and TikTok Automation Playbook — are technical operator-grade manuals for AI-augmented work. The Anti-AI Brain is the cognitive argument underneath those playbooks.

No. Claude and GPT-4o were used as sparring partners — challenging arguments, proposing counter-examples, stress-testing chapter logic, and checking citation details — but every paragraph that survived into the final manuscript was either written originally by the author or rewritten line-by-line after a model draft. The central thesis, the five-layer model, the Anti-AI Seven, the Ten-Minute Wall, the protocol structure, and the voice are author decisions on the record. Full disclosure is on the provenance page.

Amazon Kindle launched May 29, 2026. The Kindle edition is live at $9.99, and the reader bonus stack unlocks by email through the reader page. Kindle Unlimited readers can read it through KU. Paperback comes shortly.

Kindle Unlimited reach and the largest single English-language book audience on one platform. Paperback, direct sales, bundles, workbook, and the audio course come later. No plans for exclusivity past the Kindle Unlimited launch window.

Not at launch. Narration is scheduled for Q3 2026. Kindle buyers on the list will be emailed a 50%-off audiobook coupon on release.

Amazon refunds Kindle purchases within seven days of purchase with no questions asked. If you read the book and decide the protocol is not for you, Amazon will refund you under their standard Kindle return terms.