For publishers, rights holders, and author platforms
Amazon decides how your books read.
Until you take it back.
Amazon's Ask This Book launched in December 2025. Apple, Google, and others are following. Your titles are being interpreted by commercial AI defaults - and your editors have no way to govern what readers hear. Canon Pack Lite (CP-Lite and CP-Full) is the editorial-control layer publishers use to keep that authority. A free four-week pilot with one title from your list is available.
§ 01 The problem
The reading posture is being licensed too.
Right now, you don't control it.
When a reader asks AI about your book inside Amazon Kindle, Apple Books, or any platform with an AI reading layer, the answer is shaped by the model's training defaults - therapeutic, motivational, productivity-coded, smoothly reassuring. Applied to a Stoic notebook, a civic autobiography, or a serious work of nonfiction, that voice is not factually wrong. It is wrong in mode.
We call this interpretive register drift. At platform scale, it silently rewrites what your books mean to your readers.
- Editorial brand erosion at platform scale.
- Author dissatisfaction, with no recourse.
- Zero differentiation from generic AI on every other publisher's titles.
- No control over the layer your readers spend more and more of their time in.
§ 02 The difference, shown
Same question. Two voices.
One is your edition's. One isn't.
"Marcus repeatedly tells himself to stop caring what others think. What does that repetition tell us?"
Generic commercial model
Marcus's repeated reminders show us the importance of self-affirmation. By returning to the same lessons, he was practicing a form of mindfulness that helped him stay grounded. This is a powerful reminder that our values need daily reinforcement - much like how we use affirmations or journaling today to keep our priorities front and center. Would you like some practical ways to apply this in your own life?
Your title under CP-Lite governance
The repetition is doing work the Stoic project demands. Marcus is not affirming - he is correcting. The same lesson returns because the deviation returns. Read closely, the Meditations is less a notebook of insight than a record of relapse: a record of how often the trained mind drifts from what it has already understood. The repetition is the philosophy under operating conditions, not its decoration.
Neither response is hallucinated. The difference is interpretive register — the editorial voice your readers encounter when they engage with your title through AI.
§ 03 Five paths into AI-mediated reading
Only one keeps editorial control with you.
| Approach | What governance you actually retain |
|---|---|
| Generic AI reading layer (Ask This Book and similar) | None. Default model behavior, no editorial input, no opt-out. |
| Custom GPT or prompt template | Limited. Soft instructions easily overridden, no preview, no enforcement, no audit trail. |
| Fine-tuned vendor model | Partial. Expensive, slow to update, opaque to editors, requires platform cooperation. |
| In-house build from scratch | Theoretically full. Practically, 6–18 months and seven-figure cost before first title ships. |
| Canon Pack Lite (CP-Lite) | Full editorial control. Inspectable, revisable, previewable. Deployable on your platform or ours. |
§ 04 Live in production
Nine canonical works, governed today.
The reference implementation has been running on authors.living-literature.org since February 2026, with public titles spanning Stoic philosophy, civic autobiography, theological confession, strategic doctrine, slave narrative, and dramatic literature. Every title was onboarded through the same Canon Pack methodology described above. Browse the live shelf to interrogate any of them directly.
Send us one title. We'll return a governed companion in four weeks.
Free, no obligation. If you continue, subsequent titles use per-title pricing. If you don't, you keep the report and the workspace data.