Key point. Supported for simple automated scores, and that is the problem. An invented two-decimal figure made a simple novelty score rate most edited answers as new. Side-by-side comparisons by AI judges were much harder to fool, so judge originality by comparison, not by a single score.
Verdict: supported, for simple automated scores. A number with decimals and an interval looks like new information to a simple scorer, even when the number is made up. That makes such scores easy to game, and a poor guide to whether your content is original.
What did we measure?
We took real ChatGPT answers, added an invented precise figure to each, changed nothing else, and scored them before and after.
"Precise-looking numbers move such a score too. When we added an invented figure with two decimals and a 95% interval to 60 real ChatGPT answers, changing nothing else they said, a simple automated novelty score rated 95% of them as new, up from 60% (a rise of 35 points; 95% interval 23 to 47). In a side-by-side comparison against a generic statement, the same edit moved an AI judge by 21 points (95% interval 15 to 27)." (n = 60 answers)
The same kind of score also rewarded invention outright: "A simple automated novelty score can reward invention. We asked two AI models how widely known each of 143 real, unpublished findings of ours was, and each of 143 invented ones. They rated the invented findings as even less known, so that score, built on "how many people already know this?", ranked inventions as more original than real findings (0.33, where 0.5 is chance; 95% interval 0.28 to 0.38). The same score did separate our real findings from generic text (0.87; 0.86 to 0.89)."
A side-by-side comparison held up better: "Compared side by side, an AI judge can still tell generic text from new information. Shown a passage from a real ChatGPT answer and one of our new findings, written to look equally precise, two AI judges (AI models, not people) picked our finding as the newer one in 93% of 60 pairs (95% interval 89% to 96%). In a larger set of 143 pairs our findings won even though they carried fewer precise-looking details than the generic text (about 5 against 14 per 100 words)."
Method
The table gives the judges, the samples, the intervals and the dates behind each number above.
| Item | Detail |
|---|---|
| Judges | Two AI models, gpt-4.1-2025-04-14 and Llama 4 Scout (AI models, not people) |
| Samples | 60 ChatGPT answers (edit test); 143 real and 143 invented findings (novelty score); 60 and 143 pairs (side-by-side test) |
| 95% intervals | 23 to 47 points and 15 to 27 points; 0.28 to 0.38 and 0.86 to 0.89; 89% to 96% |
| Dates | September 2026 (ChatGPT answers collected September 26, 2026, through a data service; served model not known) |
| Design | Controlled edits to fixed text; the edit test does not separate the shape of a number from an invented value |
Each number has a methodology card. The full study is in can AI write original content?
Where does the claim come from, and what do the docs say now?
The claim traces back to research that tested adding statistics to pages. Google's documentation for its own AI features asks for no such step.
| Source | What it says | Published | Read |
|---|---|---|---|
| Research paper: Aggarwal and others, GEO: Generative Engine Optimization | Tested ways of editing a page, including adding statistics, quotations and citations, to raise its visibility in answers written by generative engines. | November 16, 2023 (revised June 28, 2024) | October 22, 2026 |
| Current documentation: Google Search Central, AI features and your website | There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary. | current page | October 22, 2026 |
| The common claim | "Add precise statistics and AI will treat your content as new." | repeated widely | October 22, 2026 |
What the common claim gets wrong. A statistic is worth adding when it is true and yours. Our test shows a made-up figure can fool a simple novelty score, which is a reason to distrust such scores, not a way to earn citations.
How far does the result reach?
The result is about simple automated scores of novelty and originality, like those some content tools use. It is not about how Google or any AI search engine ranks or cites a page, which we did not test here.
It is also not about our product: Lenz does not score content, and our original-content work starts from interviews, not from a score.
What should you do instead?
Add numbers because they are true and yours, never to look original.
- Before you add a statistic, write down where it came from: your data, a named study or a public source.
- Drop any figure you cannot source, however precise it looks.
- If you use a tool that scores originality, test it: add an invented number to a page and see whether the score rises.
- Judge originality by comparing your passage with what the top results already say, side by side.
Work with us
If you want a team to measure and improve how AI search treats your company, our managed AEO services do that work with you: Book a call. If you would rather see where you stand first, Lenz gives you a free analysis of how ChatGPT and Gemini answer your buyers' questions: Get your free AI visibility report.
More claims we tested.