---
title: "Know Your Audience — paper by Rooein et al. (2023)"
description: "The paper asks whether large language models can tailor their answers to readers of different ages… 17 references, cited by 11 papers here."
authors: ["Donya Rooein", "Amanda Cercas Curry", "Dirk Hovy"]
year: 2023
venue: "arXiv:2312.02065"
source: "https://arxiv.org/abs/2312.02065"
doi: "10.48550/arxiv.2312.02065"
arxiv: "2312.02065"
licence: "CC BY 4.0 (arXiv)"
licence_url: "http://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/rooein-2023-know-audience-llms-adapt-different/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/rooein-2023-know-audience-llms-adapt-different/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/11/) / Know Your Audience: Do LLMs Adapt to Different Age and Education Levels?

# Know Your Audience: Do LLMs Adapt to Different Age and Education Levels?

**Authors:** [Donya Rooein](https://md.research.elamur.ai/free/en/authors/donya-rooein/), [Amanda Cercas Curry](https://md.research.elamur.ai/free/en/authors/amanda-cercas-curry/) and [Dirk Hovy](https://md.research.elamur.ai/free/en/authors/dirk-hovy/) · **Year:** 2023 · **Venue:** arXiv:2312.02065 · **Licence:** [CC BY 4.0 (arXiv)](http://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2312.02065) · [PDF](https://arxiv.org/pdf/2312.02065) · [DOI 10.48550/arxiv.2312.02065](https://doi.org/10.48550/arxiv.2312.02065) · [arXiv 2312.02065](https://arxiv.org/abs/2312.02065) · [OpenAlex](https://openalex.org/W4389364473)

**Citations:** 2 citations (OpenAlex, 2026-10-14) · [11](https://md.research.elamur.ai/free/en/papers/rooein-2023-know-audience-llms-adapt-different/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary

The paper asks whether large language models can tailor their answers to readers of different ages and education levels, which matters as these models enter classrooms. The authors had ChatGPT produce a pool of science questions, then selected 100 that span fields such as physics, biology, chemistry and math. They built prompt templates naming a target audience: a specific age, a US school grade, a difficulty level, or a broad group such as kids, adults or professionals. Four models answered these prompts over repeated runs: two commercial (ChatGPT and GPT-3 text-davinci-003) and two open-source (bigscience-T0 and Flan-T5). Each answer was scored with the Flesch-Kincaid Reading Ease (FKRE) and Grade Level (FKGL) metrics. The scores were then compared with the readability range recommended for the requested audience.

The models turned out to have fairly fixed readability profiles that prompting barely changes. The commercial models consistently wrote at an advanced level, while the open-source ones produced scattered scores. The authors conclude that current LLMs are poorly suited to audience-adapted educational use. They also caution that standard readability formulas may not fully fit machine-generated text.

## Key points

- **Corpus.** The dataset contains 33,600 prompt–response pairs built from 100 science questions. It is released together with the computed readability metrics.
- **Difficulty scale.** The FKRE score was mapped to seven difficulty bands. For the analysis, these were merged into three: easy (71–100), medium (60–70) and hard (0–59).
- **ChatGPT.** Its answers fell in the correct FKRE range almost only for the most advanced targets:
  - 100% for 22–23-year-olds,
  - 94% for college graduates,
  - 99% for "difficult-level" prompts,
  - essentially 0% for younger readers.
- **Other models.** GPT-3 reached 60% in-range for 18–19-year-olds and 56% for college students. Bigscience-T0 and Flan-T5 spread their answers across many ranges. Average in-range rates were similar for all four models, at roughly 14.6–15.7%.
- **Vocabulary diversity and limits.** Type-Token Ratio stayed almost flat across targets; for ChatGPT it was 0.637 for an 11-year-old versus 0.635 for 22–23-year-olds. The authors note that FKRE and FKGL are designed for English and the US school system.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Bocconi University](https://md.research.elamur.ai/free/en/universities/bocconi-university/) — University, Milan, Italy · 28 works · 3,077 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,437 works

🧭 Topics (1): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works

## Cite

Cite as: Donya Rooein, Amanda Cercas Curry and Dirk Hovy. 2023. Know Your Audience: Do LLMs Adapt to Different Age and Education Levels?. arXiv:2312.02065.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{rooein2023know,
  title = {{Know Your Audience: Do LLMs Adapt to Different Age and Education Levels?}},
  author = {Donya Rooein, Amanda Cercas Curry and Dirk Hovy},
  year = {2023},
  doi = {10.48550/arxiv.2312.02065},
  eprint = {2312.02065},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2312.02065}
}
```

- [← previous Instruction Clarification Requests in Multimodal Collaborative Dialogue Games: Tasks, and an Analysis of the CoDraw Dataset](https://md.research.elamur.ai/free/en/papers/madureira-2023-instruction-clarification-requests-multimodal-collaborative/)
- [next → Retrieval-augmented Multi-modal Chain-of-Thoughts Reasoning for Large Language Models](https://md.research.elamur.ai/free/en/papers/liu-2023-retrieval-augmented-multi-modal-chain/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/11/)

## References

17 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Tarek AIT BAHA, Mohamed EL HAJJI, Youssef ES-SAADY, and Hammou FADILI. 2022. Towards highly adaptive edu-chatbot. Procedia Computer Science, 198:397–403. 12th International Conference on Emerging Ubiquitous Systems and Pervasive Networks / 11th International Conference on Current and Future Trends o… [find](https://scholar.google.com/scholar?q=Tarek+AIT+BAHA%2C+Mohamed+EL+HAJJI%2C+Youssef+ES%2DSAADY%2C+and+Hammou+FADILI%2E+2022%2E+Towards+highly+adaptive+edu%2Dchatbot%2E+Procedia+Computer+Science%2C+198%3A397%E2%80%93403%2E+12th+International+Conference+on+Emerging+Ubiq)
2. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. [Language models are few-shot learners](https://md.research.elamur.ai/free/en/papers/brown-2020-language-models-few-shot-learners/). Advances in neural information processing systems, 33:1877–1901.
3. Sadia Choudhery, Yin Xi, Heng Chen, Nader Aboul-Fettouh, YPaul Goldenmerry, Charles Ho, Hannah Viroslav, Chelsea Zhang, and Sally Goudreau. 2020. Readability and quality of online patient education material on websites of breast imaging centers. Journal of the American College of Radiology, 17(10):1… [find](https://scholar.google.com/scholar?q=Sadia+Choudhery%2C+Yin+Xi%2C+Heng+Chen%2C+Nader+Aboul%2DFettouh%2C+YPaul+Goldenmerry%2C+Charles+Ho%2C+Hannah+Viroslav%2C+Chelsea+Zhang%2C+and+Sally+Goudreau%2E+2020%2E+Readability+and+quality+of+online+patient+education+ma)
4. Wilson Cheong Hin Hong. 2023. The impact of chatgpt on foreign language teaching and learning: opportunities in education and research. Journal of Educational Technology and Innovation, 5(1). [find](https://scholar.google.com/scholar?q=Wilson+Cheong+Hin+Hong%2E+2023%2E+The+impact+of+chatgpt+on+foreign+language+teaching+and+learning%3A+opportunities+in+education+and+research%2E+Journal+of+Educational+Technology+and+Innovation%2C+5%281%29%2E)
5. Enkelejda Kasneci, Kathrin Seßler, Stefan Küchemann, Maria Bannert, Daryna Dementieva, Frank Fischer, Urs Gasser, Georg Groh, Stephan Günnemann, Eyke Hüllermeier, et al. 2023. [Chatgpt for good? on opportunities and challenges of large language models for education](https://md.research.elamur.ai/free/en/papers/kasneci-2023-chatgpt-good-opportunities-challenges-large/). Learning and Individual Difference…
6. Michael R King and ChatGPT. 2023. A conversation on artificial intelligence, chatbots, and plagiarism in higher education. Cellular and Molecular Bioengineering, 16(1):1–2. [find](https://scholar.google.com/scholar?q=Michael+R+King+and+ChatGPT%2E+2023%2E+A+conversation+on+artificial+intelligence%2C+chatbots%2C+and+plagiarism+in+higher+education%2E+Cellular+and+Molecular+Bioengineering%2C+16%281%29%3A1%E2%80%932%2E)
7. Louis Martin, Angela Fan, Éric De La Clergerie, Antoine Bordes, and Benoît Sagot. 2020. [Muss: multilingual unsupervised sentence simplification by mining paraphrases](https://md.research.elamur.ai/free/en/papers/martin-2022-muss-multilingual-unsupervised-sentence-simplification/). arXiv preprint arXiv:2005.00352.
8. Emiliana Murgia, Maria Soledad Pera, Monica Lan-doni, and Theo Huibers. 2023. Children on chatgpt readability in an educational context: Myth or opportunity? In Adjunct Proceedings of the 31st ACM Conference on User Modeling, Adaptation and Personalization, pages 311–316. [find](https://scholar.google.com/scholar?q=Emiliana+Murgia%2C+Maria+Soledad+Pera%2C+Monica+Lan%2Ddoni%2C+and+Theo+Huibers%2E+2023%2E+Children+on+chatgpt+readability+in+an+educational+context%3A+Myth+or+opportunity%3F+In+Adjunct+Proceedings+of+the+31st+ACM+Con)
9. Chinh Duc Nguyen. 2021. The construction of age-appropriate pedagogies for young learners of english in primary schools. The Language Learning Journal, 49(1):13–26. [find](https://scholar.google.com/scholar?q=Chinh+Duc+Nguyen%2E+2021%2E+The+construction+of+age%2Dappropriate+pedagogies+for+young+learners+of+english+in+primary+schools%2E+The+Language+Learning+Journal%2C+49%281%29%3A13%E2%80%9326%2E)
10. Riaz Qureshi, Daniel Shaughnessy, Kayden AR Gill, Karen A Robinson, Tianjing Li, and Eitan Agai. 2023. Are chatgpt and large language models “the answer” to bringing us closer to systematic review automation? Systematic Reviews, 12(1):72. [find](https://scholar.google.com/scholar?q=Riaz+Qureshi%2C+Daniel+Shaughnessy%2C+Kayden+AR+Gill%2C+Karen+A+Robinson%2C+Tianjing+Li%2C+and+Eitan+Agai%2E+2023%2E+Are+chatgpt+and+large+language+models+%E2%80%9Cthe+answer%E2%80%9D+to+bringing+us+closer+to+systematic+review+aut)
11. D. Rooein and P. Paolini. 2023. Talkytutor: An innovative approach for chatbots for education. In INTED2023 Proceedings, 17th International Technology, Education and Development Conference, pages 3088–3093. IATED. [find](https://scholar.google.com/scholar?q=D%2E+Rooein+and+P%2E+Paolini%2E+2023%2E+Talkytutor%3A+An+innovative+approach+for+chatbots+for+education%2E+In+INTED2023+Proceedings%2C+17th+International+Technology%2C+Education+and+Development+Conference%2C+pages+3088)
12. Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Eric Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, et al. 2022. Scaling instruction-finetuned language models. arXiv preprint arXiv:2210.11416. [find](https://scholar.google.com/scholar?q=Hyung+Won+Chung%2C+Le+Hou%2C+Shayne+Longpre%2C+Barret+Zoph%2C+Yi+Tay%2C+William+Fedus%2C+Eric+Li%2C+Xuezhi+Wang%2C+Mostafa+Dehghani%2C+Siddhartha+Brahma%2C+et+al%2E+2022%2E+Scaling+instruction%2Dfinetuned+language+models%2E+arXi)
13. Donya Rooein and Paolo Paolini. 2021. Adaptive conversations for adaptive learning talkytutors: sustainable chatbots as tutors. In EdMedia+ Innovate Learning, pages 1020–1029. Association for the Advancement of Computing in Education (AACE). [find](https://scholar.google.com/scholar?q=Donya+Rooein+and+Paolo+Paolini%2E+2021%2E+Adaptive+conversations+for+adaptive+learning+talkytutors%3A+sustainable+chatbots+as+tutors%2E+In+EdMedia%2B+Innovate+Learning%2C+pages+1020%E2%80%931029%2E+Association+for+the+Adva)
14. Debby RE Cotton, Peter A Cotton, and J Reuben Shipway. Chatting and cheating: Ensuring academic integrity in the era of chatgpt. [find](https://scholar.google.com/scholar?q=Debby+RE+Cotton%2C+Peter+A+Cotton%2C+and+J+Reuben+Shipway%2E+Chatting+and+cheating%3A+Ensuring+academic+integrity+in+the+era+of+chatgpt%2E)
15. Victor Sanh, Albert Webson, Colin Raffel, Stephen H Bach, Lintang Sutawika, Zaid Alyafeai, Antoine Chaffin, Arnaud Stiegler, Teven Le Scao, Arun Raja, et al. 2021. [Multitask prompted training enables zero-shot task generalization](https://md.research.elamur.ai/free/en/papers/sanh-2021-multitask-prompted-training-enables-zero/). arXiv preprint arXiv:2110.08207.
16. Hana L Haver, Cheng Ting Lin, Arlene Sirajuddin, Paul H Yi, and Jean Jeudy. 2023. Use of chatgpt, Zijian Wang, Scott Hale, David Ifeoluwa Adelani, Przemyslaw Grabowicz, Timo Hartman, Fabian Flöck, and David Jurgens. 2019. Demographic inference and representative population estimates from multilingua… [find](https://scholar.google.com/scholar?q=Hana+L+Haver%2C+Cheng+Ting+Lin%2C+Arlene+Sirajuddin%2C+Paul+H+Yi%2C+and+Jean+Jeudy%2E+2023%2E+Use+of+chatgpt%2C+Zijian+Wang%2C+Scott+Hale%2C+David+Ifeoluwa+Adelani%2C+Przemyslaw+Grabowicz%2C+Timo+Hartman%2C+Fabian+Fl%C3%B6ck%2C+and)
17. JD Zamfirescu-Pereira, Richmond Y Wong, Bjoern Hartmann, and Qian Yang. 2023. Why johnny can’t prompt: how non-ai experts try (and fail) to design llm prompts. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, pages 1–21. [find](https://scholar.google.com/scholar?q=JD+Zamfirescu%2DPereira%2C+Richmond+Y+Wong%2C+Bjoern+Hartmann%2C+and+Qian+Yang%2E+2023%2E+Why+johnny+can%E2%80%99t+prompt%3A+how+non%2Dai+experts+try+%28and+fail%29+to+design+llm+prompts%2E+In+Proceedings+of+the+2023+CHI+Conferenc)

*Updated 31 October 2026*