On Seeding Watermarks to Detect Verbatim LLM Copy-Paste Responses
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Security Radio. Generated commentary on the latest security and cryptography papers.
Nadia: We're looking at the paper: "On Seeding Watermarks to Detect Verbatim LLM Copy-Paste Responses".
Elias: Here's the gist of the paper. Large language models (LLMs) have made fluent essay writing, code drafting, and quiz answering instantly available to students at every level from secondary school through graduate study <ref:2605.16336#pg2>.
Nadia: Many educators do not object to LLM use per se; what they need to detect is the case in which a student pastes the assignment prompt into a chatbot and submits the model’s reply verbatim without engaging with the work <ref:2605.16336#pg3>.
Elias: The paper proposes SteganoPrompt, an inputside watermark embedded inside the visible assignment prompt itself, which an LLM reads to write a tell-tale signature into its reply specifically exposing the copy-and-paste pathway <ref:2605.16336#pg2>.
Nadia: The Problem with Existing Detection Methods Existing posthoc AI-text detectors remain unreliable and have been shown to penalise non-native English writers, while output-side watermarks require cooperation from the model provider <ref:2605.16336#pg3>.
Paper summary: Nadia: So we're talking about this paper called "On Seeding Watermarks to Detect Verbatim LLM Copy-Paste Responses." Essentially, they’re looking at how students are using AI to write essays and the problem is that teachers can't easily tell when a student just pastes the whole assignment prompt into a chatbot and turns in the AI's reply without actually doing any of the work >
Elias: Exactly. The paper says existing methods for spotting this, like those looking at text after it’s been written, aren't very reliable because they can end up punishing students who write in a way that isn't standard English or non-native speakers have >
Priya: And output watermarks are another problem because you need the model provider to cooperate, which most teachers don't have control over for their classrooms >
Nadia: So what they propose is something different. They suggest an inputside watermark, meaning they embed an invisible instruction directly into the assignment prompt itself, and if an AI reads that prompt verbatim it writes a secret signature into its response to expose the copy-paste pathway >
Elias: That specific technique is called SteganoPrompt, and it uses a method where they encode something printable in ASCII into those Unicode tags block U+E0000 to U+E007F which looks exactly like the original text >
Nadia: It sounds like they're trying to create a channel that the student doesn't even see, so if an AI ingests the prompt verbatim, it quietly reads that hidden instruction and writes a tell-tale signature into its reply >
Elias: The method involves setting A equal to Ap union Aw and defining T as c plus zero E000 colon where C is in A >
Priya: I'm wondering about the practical side of this. If the goal is to detect verbatim copying, how does this inputside approach actually survive the process of a student using it? >
Nadia: Well, they focused on making sure it has visual identity and that it survives common distribution pipelines, which means testing if it gets stripped out by different text-editing tools >
Elias: They found that models like Claude and Gemini emitted this integrity token in every trial, but GPT series showed none because of something called input scrubbing >
Priya: That's interesting about the model compliance. But what about the channel survival part? Does it actually stay intact when people copy that text to different places? >
Nadia: They found that the payload is preserved through every text-editing pipeline they tested, with only some social platforms stripping extended Unicode characters >
Elias: They also mentioned linguistic steganography as another area, like using context-aware lexical substitution to trace where a passage came from, but they noted those methods change the wording which isn't what you want for an assignment brief that needs to stay verbatim >
Priya: So the Unicode tag approach is different because it modifies the byte stream while leaving every rendered glyph unchanged, which keeps the text looking normal >
Nadia: Right. And they built this tool as a single page web tool that does all the computation right in your browser, which makes it pretty accessible for someone trying to test this out >
Elias: The recommended way to use it involves four steps: first disclosing the policy in the syllabus, second encoding each brief with SteganoPrompt and pasting it into the learning management system, third verifying before releasing by decoding the brief to confirm the payload is bitidentical, and finally searching submissions for that literal token at grading time >
Priya: I wonder about that verification step. If you decode it to check for a bitidentical payload, does that really give you confidence in catching someone who might be slightly altering the prompt? >
Nadia: The paper argues that this workflow generalizes the manual word of the day check without needing visible student action, which is a big deal for teachers trying to keep up with this AI use >
Elias: Beyond just detecting verbatim copying, they framed it as an alternative that you control directly, sitting between output watermarking and indirect prompt injection attacks >
Priya: What are the ethical considerations here? The paper touches on the need to disclose the policy to have a deterrent effect and using a kind payload that doesn't embarrass students if they make a mistake >
Nadia: They suggest that if someone hits this watermark, it should be treated as a starting point for conversation instead of an immediate verdict on their integrity >
Elias: And they are careful about using the technique against systems you don't own or haven't been authorized to test because it can be used in adversarial ways, like indirect prompt injection >
Priya: So, what’s the actual future direction for this work? What do they suggest doing next after proving this inputside watermark works for detecting copy-paste? >
Nadia: They suggested embedding an HMAC over assignment metadata to get stronger verification of the assignment itself >
Elias: They also proposed combining that Unicode-Tag channel with a content level watermark to help it survive light paraphrasing, and they plan a multi section, multi instructor field study to estimate the actual detection rates across different scenarios >
Priya: That sounds like a solid plan for getting more robust data on how effective this is in the real world, especially when dealing with different types of student work >
Conclusion: Nadia: So we're wrapping up this talk on "On Seeding Watermarks to Detect Verbatim LLM Copy-Paste Responses." Basically, the paper shows an inputside way to watermark assignment prompts so if an AI copies it verbatim, it leaves a traceable signature in its answer.
Elias: It’s about embedding something into the prompt itself that the AI reads and then writes back. The authors used something called SteganoPrompt which uses those unused Unicode tags block U plus E zero zero zero to U plus E zero seven f which looks exactly like normal text.
Priya: So, what does this mean for us? We're talking about making sure students aren't just pasting the whole question into a chatbot and submitting the AI's answer without doing the actual thinking.
Nadia: Right, it’s a direct way to check that specific copying behavior without having to look at the final essay text. The authors tested this across different language models like Claude and Gemini, and they saw that most of them put out this integrity token in every test run.
Elias: But the results weren't perfect. They found that some models, specifically a GPT series model, didn't show any token because it had some input scrubbing mechanism. That’s a big caveat for us to keep in mind when we talk about model compliance across the board.
Priya: So, even with this watermark on the prompt side, you still have to be careful about which AI you are testing against and how that specific model handles instructions before you can trust the detection. The data shows that channel survival is good; they found it stays intact through most text editing pipelines.
Nadia: It’s a tool sitting right between trying to detect what the AI writes and trying to control what the AI produces at output time. It gives educators a way to catch that direct copy-paste action during grading, which is currently pretty hard because of all the paraphrasing students do.
Elias: Exactly. The authors suggest this inputside watermark is a working classroom integrity tool, but they also point out it’s not perfect on its own. They plan more work to combine this Unicode channel with a content level watermark to handle light paraphrasing better.
Priya: And they’re looking at embedding something called an HMAC over assignment metadata for even stronger verification of the assignment itself before you even get to the student submissions. That suggests a layered approach might be needed here.
Nadia: So, this paper gives us a concrete method—SteganoPrompt—that allows instructors to search for that literal token at grading time, which is a big step toward making academic integrity checks more practical and less reliant on just looking at the final paper.
George Washington University
cs.CR, cs.AI, cs.CL, cs.CY, cs.ET
Submitted: 2026-05-07
Updated: 2026-08-07
Comments: We are pleased to announce that this paper has been accepted by the 9th AAAI/ACM Conference on AI, Ethics, and Society (AIES 2026). We appreciate the valuable feedback from the reviewers and look forward to sharing our findings with the community
Journal ref: Proceedings of the Ninth AAAI/ACM Conference on AI, Ethics, and Society (AIES 2026)
Code: https://github.com/Ezharjan/SteganoPrompt
Project page: https://ezharjan.github.io/SteganoPrompt
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 82/100
The gist: Large language models (LLMs) have made fluent essay writing, code drafting, and quiz answering instantly available to students at every level from secondary school through graduate study…
Key concepts
- SteganoPrompt
- This is an inputside watermarking technique where a hidden, printable-ASCII message is encoded directly into the visible text of an assignment prompt using specific Unicode blocks. The goal is to leave the text visually unchanged while embedding a secret signal for detection.
- Inputside Watermarking
- Instead of marking the output, this method modifies the input (the prompt) before it reaches the LLM. SteganoPrompt achieves this by encoding data into unused Unicode tags, making it an 'invisible' modification that triggers a specific response from the model.
- Channel Survival
- This refers to how well the embedded watermark survives common text processing steps, like copying and pasting or passing through different text-editing pipelines. The technique is designed to ensure the hidden payload remains intact even when the prompt is manipulated by students.
- Cross-Model Compliance
- This measures whether different large language models (like Claude or Gemini) correctly recognize and respond to the hidden instruction embedded in the SteganoPrompt watermark. The paper found that some models respected these hidden instructions while others did not.
Terminology
Summary
Large language models (LLMs) have made fluent essay writing, code drafting, and quiz answering instantly available to students at every level from secondary school through graduate study <ref:2605. p. 0 On Seeding Watermarks to Detect Verbatim LLM Copy-Paste Responses
Aizierjiang Aiersilan Q Artin Yousefi Robert Pless
The George Washington University
{alexandera, artinyousefi, pless}@gwu. edu
Abstract
Large language models (LLMs) have made fluent essay writing, code drafting, and quiz answering instantly available to
students at every level, from secondary school through graduate study. edu
Abstract
Large language models (LLMs) have made fluent essay writing, code drafting, and quiz answering instantly available to
students at every level, from secondary school through graduate study. Many educators do not object to LLM use per se;
what they need to detect is the case in which a student pastes
the assignment prompt into a chatbot and submits the model’s
reply verbatim, without engaging with the work. We propose an alternative that the educator controls directly: an inputside watermark in which an invisible instruction is embedded
inside the visible assignment prompt itself. An LLM that ingests the prompt verbatim quietly reads the hidden instruction
and writes a tell-tale signature into its reply, exposing the copyand-paste pathway specifically.Where the paper says it
Where the paper says it
Where the paper says it
The Problem with Existing Detection Methods
Existing posthoc AI-text detectors remain unreliable and have been shown to penalise non-native English writers, while output-side watermarks require cooperation from the model provider <ref:2605.16336#pg3>. The paper identifies three options for countermeasures: post-hoc detection, output-side watermarking, and inputside watermarking <ref:2605. p. 1 Taken together, post-hoc
detection is informative as a signal but not as a verdict. Output-Side Watermarking of LLM Text
A second, complementary line embeds a watermark at the
model end. 2023; Krishna et al. 2023) results have shown that such detectors are unreliable,
can be defeated by light paraphrasing, and disproportionately
misclassify essays written by non-native English speakers. 2023; Christ, Gunn, and Zamir 2024; Zhao et al. Output watermarking is technically attractive but is only
available to parties who control the model, while classroom
teacher or university instructor generally does not.Where the paper says it
Where the paper says it
Where the paper says it
How the Training Signal is Built
SteganoPrompt encodes an arbitrary printable-ASCII payload into the deprecated Unicode Tags block U+E0000–U+E007F, which is visually identical to the original text <ref:2605.16336#pg2>. The encoding procedure involves setting A = Ap ∪ Aw and defining T = c + 0xE000: c ∈ A (1) <ref:2605. p. 2 The instructor must be able to decode any
text and inspect the payload before distribution, so that
the hidden message contains exactly what was intended. Encoding Procedure
Let Ap = { c ∈ N : 0x20 ≤ c ≤ 0x7E } denote the 95
printable ASCII code points, and let Aw = {0x09, 0x0A}
denote the horizontal tab and the line feed, which we carry
through encoding so that the layout of multi-line payloads is
preserved. \n is encoded
as
E(W) = ⟨ U+E0048, U+E0069,
U+E0021, U+E000A ⟩,
none of which renders as a glyph in any common font, but
each of which is tokenized by current chatbots as the corresponding ASCII byte. Given a visible text P ∈ A∗
and a payload W ∈ A∗
, the
watermarked document is the concatenation
wmπ(P, W) = P[1:k] ∥ E(W) ∥ P[k+1:|P |]
, (4)
where ∥ denotes string concatenation, P[i:j]
is the substring
of P from position i to position j (and is empty if j < i), , |P|} depends on the
placement parameter π. Define the first-sentence-boundary
index
µ(P) = min
j : Pj ∈ {. An LLM that ingests the prompt verbatim quietly reads the hidden instruction
and writes a tell-tale signature into its reply, exposing the copyand-paste pathway specifically. We describe SteganoPrompt,
a single-page, zero-dependency web tool that encodes an arbitrary printable-ASCII payload into the deprecated Unicode
Tags block (U+E0000–U+E007F).Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the Gains Come From
The technique is designed around five goals, including visual identity, channel survival, reader transparency, and privacy and offline use <ref:2605.16336#pg5>. The evaluation focuses on two technical axes: cross-model compliance with the hidden instruction and cross-channel survival of the payload through common distribution pipelines <ref:2605. p. 4 The visible assignment
text is never altered, and the Verify panel lets the instructor
confirm the exact payload before distribution. Evaluation
We evaluate SteganoPrompt on two technical axes that
are controllable independently of the student population:
(i) cross-model compliance with the hidden instruction, and
(ii) cross-channel survival of the payload through common
distribution pipelines. The latest GPT-5-
series models appear to apply such a filter, and independent
audits report the same input scrubbing in some other hosted
chat products (Markopoulos and Fagan 2025; Toulas 2025). Other providers have not adopted it: Google has stated that
it does not treat the channel as a security issue for Gemini
(Toulas 2025), and the Claude, Grok, Qwen, DeepSeek, and
Codestral pipelines all delivered the Tag block to the model
unfiltered. Cross-Channel Survival
Table 2 summarizes the channel-survival study. The payload
is preserved through every text-editing pipeline we tested, including Microsoft Word (.Where the paper says it
Where the paper says it
Where the paper says it
Recommended Deployment Workflow
The recommended workflow for an instructor involves four steps: 1) Disclose the policy in the syllabus <ref:2605.16336#pg8>; 2) Encode each brief using SteganoPrompt and paste it into the LMS <ref:2605. p. 6 Encode each brief. Paste the visible assignment text into
SteganoPrompt, choose (or write) a payload, and paste
the encoded result into the learning-management system. Verify before release. Use the Verify panel to decode
the brief from the same channel a student would read it
through, and confirm that the recovered payload is bitidentical to the intended one. Search at grading time. Search submissions for the literal token (by default, SteganoPrompt-OK-2026). The deterrent effect
of an announced policy is, by all available evidence on classroom integrity, larger than the catch effect on any individual
assignment, and disclosure is also the ethically required posture. The workflow generalizes the manual word-of-the-day
check that originally motivated this work, without requiring
the student to do anything visible: the tripwire travels with
the prompt rather than with the student.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Ethics and Disclosure
The responsible deployment of input-side watermarks has four properties, including disclosing the policy to maximize deterrent effect <ref:2605.16336#pg9>. The paper argues for using a kind payload that helps students in good faith, avoiding content that would embarrass them <ref:2605. p. 7 Use a kind payload. Prefer payloads that, when followed,
help a student who is using AI in good faith. Treat hits as a starting point. A watermark hit is a
probable-cause signal for a conversation, not a verdict. The same technique is, in adversarial hands, an indirect prompt injection. We do not endorse its use against any system the
instructor does not own or has not been authorized to test.Where the paper says it
Where the paper says it
Where the paper says it
Conclusion and Future Work
The authors present SteganoPrompt as a working classroom integrity tool that sits between output-side watermarking and indirect prompt injection <ref:2605.16336#pg10>. Future work suggests embedding an HMAC over assignment metadata for stronger verification <ref:2605. p. 8 Several directions are worth pursuing. First, the current
payload is a literal string; a stronger design would embed an
Hash-based Message Authentication Code (HMAC) over the
assignment metadata, so that the instructor could verify not
only that a watermark is present but that it is the watermark
associated with this assignment, this term, and this section. 2022; Ueoka,
Murawaki, and Kurohashi 2021), would survive a student
who lightly paraphrases the brief before pasting it. Third,
a pre-registered, multi-section, multi-instructor field study
would allow detection and deterrence rates to be estimated
with statistical confidence; we did not run such a study and
we leave it to follow-on work. Fourth, as model pipelines
incorporate input sanitizers, the design space for integrity
watermarks will narrow; we are interested in a principled
negotiation between the educational community and model
providers on which input transformations should and should
not be applied silently to user-pasted text. The tool is released under the MIT licence at https:
//ezharjan. This work was motivated by two recent observations. First,
the program chairs of ICML 2026 announced that reviewerfacing materials would be circulated with hidden watermarks,
so that a reviewer who pasted a call into an LLM and copied
the model’s reply back into the review form could be flagged
arXiv:2605.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
References
Aaronson, S.; and Kirchner, H. p. 0 The second family
is output-side watermarking, in which the model provider
biases its sampling distribution at generation time so that a
downstream verifier can later test for the watermark (Kirchenbauer et al. 2023; Aaronson and Kirchner 2023; Kuditipudi
et al. ; and Kirchner, H. Watermarking of large
language models.Where the paper says it
Where the paper says it
Bommasani, R. p. 0 com/Ezharjan/SteganoPrompt
Introduction
The release of broadly capable instruction-following language models (Ouyang et al. 2022; Bommasani et al. Bommasani, R. ; Hudson, D. ; Hudson, D. ; Adeli, E. ; Adeli, E. ; Altman, R. ; Altman, R. ; Arora,
S. ; Arora,
S. ; von Arx, S. ; von Arx, S. ; Bernstein, M. ; Bernstein, M. ; Bosselut, A. ; Bosselut, A. ;
Brunskill, E. ;
Brunskill, E. On the opportunities and risks of
foundation models.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Boucher, N. p. 1 SteganoPrompt sits
within this conceptual frame, but with an inverted threat
model: the party embedding the hidden instruction is not an
attacker, but the legitimate author of the document, and the
victim of the injection is a copy-and-paste workflow that the
author wishes to discourage. Unicode and Character-Level Channels
Boucher et al. Boucher, N. ; Shumailov, I. ; Shumailov, I. ; Anderson, R. ; Anderson, R. ; and Papernot, N. ; and Papernot, N. Bad characters: Imperceptible nlp attacks.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Christ, M.; Gunn, S.; and Zamir, O. p. 0 2023; Aaronson and Kirchner 2023; Kuditipudi
et al. 2023; Christ, Gunn, and Zamir 2024; Zhao et al. ; and Zamir, O. Undetectable
watermarks for language models.Where the paper says it
Where the paper says it
Cotton, D. R.; Cotton, P. p. 0 2021)
has made it trivial for a student, whether in secondary school
or at university, to obtain a fluent answer to almost any shortanswer or short-essay assignment by pasting the prompt into
a chatbot. Surveys in education and rapid reviews of the resulting literature (Cotton, Cotton, and Shipway 2024; Susnjak
and McIntosh 2024; Lo 2023; Perkins 2023; Stokel-Walker
Copyright © 2026, Association for the Advancement of Artificial
Intelligence (www. 2021)
has made it trivial for a student, whether in secondary school
or at university, to obtain a fluent answer to almost any shortanswer or short-essay assignment by pasting the prompt into
a chatbot. Surveys in education and rapid reviews of the resulting literature (Cotton, Cotton, and Shipway 2024; Susnjak
and McIntosh 2024; Lo 2023; Perkins 2023; Stokel-Walker
Copyright © 2026, Association for the Advancement of Artificial
Intelligence (www. ; and Shipway, J. Chatting and cheating: Ensuring academic integrity in the era of
ChatGPT.Where the paper says it
Where the paper says it
Where the paper says it
Eger, S.; S¸ahin, G. p. 8 Innovations in education and teaching international, 61(2): 228–239. ; S¸ahin, G. ; S¸ahin, G. ; Ruckl ¨ e, A. ; Ruckl ¨ e, A. ; Schulz, C. ; Schulz, C. ; ´
Mesgar, M. ; ´
Mesgar, M. ; Swarnkar, K. ; Swarnkar, K. ; Simpson, E. ; Simpson, E. ; and Gurevych, I. ; and Gurevych, I. Text processing like humans do: Visually attacking
and shielding NLP systems.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Goodside, R. p. 5 Compliance is the
token-emission rate: high (every trial), medium (50–99%),
low (1–49%), none (never). Ratings combine our experiments with public vendor and security disclosures (Goodside 2024; Markopoulos and Fagan 2025; Toulas 2025; Rehberger 2024b,c; Liu 2025; Yeung 2026). Goodside, R. Invisible Prompt Injection via
Unicode Tag Characters.Where the paper says it
Where the paper says it
Greshake, K. p. 1 Indirect Prompt Injection
A separate strand in the security literature studies what happens when an LLM consumes content that the user did not
author. Greshake et al. Greshake, K. ; Abdelnabi, S. ; Abdelnabi, S. ; Mishra, S. ; Mishra, S. ; Endres, C. ; Endres, C. ;
and Fritz, M. ;
and Fritz, M. Not what you’ve signed up for: Compromising real-world llm-integrated applications with indirect
prompt injection.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Hines, K.; Lopez, G. p. 8 In Proceedings of the 16th ACM workshop
on artificial intelligence and security, 79–90. ; Lopez, G. ; Lopez, G. ; Zarfati, F. ; Zarfati, F. ; Zunger,
Y. ; Zunger,
Y. ; and Kiciman, E. ; and Kiciman, E. Defending against indirect
prompt injection attacks with spotlighting.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
ICML 2026. 2026. On Violations of LLM Review Policies. p. 8 arXiv preprint
arXiv:2403. On Violations of LLM Review Policies.Where the paper says it
Kirchenbauer, J. p. 0 This problem is likely to deepen over time, as human learners
of English are themselves exposed to and may internalize the
stylistic conventions of AI-generated text. The second family
is output-side watermarking, in which the model provider
biases its sampling distribution at generation time so that a
downstream verifier can later test for the watermark (Kirchenbauer et al. Kirchenbauer, J. ; Geiping, J. ; Geiping, J. ; Miers, I. ; Miers, I. ;
and Goldstein, T. ;
and Goldstein, T. A watermark for large language
models.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Krishna, K. p. 0 2023) and theoretical (Sadasivan et al. 2023; Krishna et al. Krishna, K. ; Karpinska, M. ; Karpinska, M. ; Wieting, J. ; Wieting, J. ; and Iyyer,
M. ; and Iyyer,
M. Paraphrasing evades detectors of ai-generated text,
but retrieval is an effective defense.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Kuditipudi, R. p. 0 The second family
is output-side watermarking, in which the model provider
biases its sampling distribution at generation time so that a
downstream verifier can later test for the watermark (Kirchenbauer et al. 2023; Aaronson and Kirchner 2023; Kuditipudi
et al. Kuditipudi, R. ; Thickstun, J. ; Thickstun, J. ; Hashimoto, T. 2023),
the (since-withdrawn) OpenAI classifier (OpenAI 2023), and
a number of commercial services. Both empirical (Liang et al. ; and Liang, P. Robust distortion-free watermarks for language models.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Liang, W.; Izzo, Z.; Zhang, Y. p. 8 arXiv preprint arXiv:2307. ; Zhang, Y. ; Huang, Z. Monitoring
ai-modified content at scale: A case study on the impactWhere the paper says it
Where the paper says it
Liang, W.; Yuksekgonul, M. p. 9 arXiv preprint
arXiv:2403. ; Yuksekgonul, M. ; Yuksekgonul, M. ; and Zou, J. ; and Zou, J. GPT detectors are biased against non-native English
writers.Where the paper says it
Where the paper says it
Where the paper says it
Liu, I. C. 2025. Invisible Prompt Injection: A Threat to AI Security. p. 9 Patterns, 4(7). Invisible Prompt Injection: A Threat to AI
Security.Where the paper says it
Liu, Y.; Jia, Y.; Geng, R.; Jia, J.; and Gong, N. p. 9 com/en us/research/25/a/
invisible-prompt-injection-secure-ai. ; and Gong, N. ; and Gong, N. Formalizing and benchmarking prompt injection attacks and
defenses.Where the paper says it
Where the paper says it
Lo, C. K. 2023. What is the impact of ChatGPT on education? p. 9 In 33rd USENIX Security Symposium (USENIX
Security 24), 1831–1847. What is the impact of ChatGPT on education? (2024) survey responses by universities and recommend
a combination of policy, assessment design, and detection. Perkins (2023) catalogues post-pandemic integrity considerations specific to LLMs, while Lo (2023) provides a rapid review of the early educational literature.Where the paper says it
Where the paper says it
Markopoulos, V. p. 5 Compliance is the
token-emission rate: high (every trial), medium (50–99%),
low (1–49%), none (never). Ratings combine our experiments with public vendor and security disclosures (Goodside 2024; Markopoulos and Fagan 2025; Toulas 2025; Rehberger 2024b,c; Liu 2025; Yeung 2026). Compliance is the
token-emission rate: high (every trial), medium (50–99%),
low (1–49%), none (never). Ratings combine our experiments with public vendor and security disclosures (Goodside 2024; Markopoulos and Fagan 2025; Toulas 2025; Rehberger 2024b,c; Liu 2025; Yeung 2026). ; and Fagan, A. Ghosts in
the Machine: ASCII Smuggling Across Various LLMs.Where the paper says it
Where the paper says it
Where the paper says it
Mitchell, E. p. 0 The first
is post-hoc detection, in which a classifier inspects the submitted text and decides whether it was produced by a model. Tools in this family include DetectGPT (Mitchell et al. Mitchell, E. ; Khazatsky, A. ; Khazatsky, A. ; Manning, C. ; Manning, C. ; and
Finn, C. ; and
Finn, C. Detectgpt: Zero-shot machine-generated text
detection using probability curvature.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
OpenAI. 2023. New AI Classifier for Indicating AIWritten Text. p. 9 In International conference on machine learning, 24950–24962. New AI Classifier for Indicating AIWritten Text.Where the paper says it
Ouyang, L.; Wu, J.; Jiang, X. p. 9 com/blog/new-ai-classifier-forindicating-ai-written-text. ; Jiang, X. ; Jiang, X. ; Almeida, D. ; Almeida, D. ; Wainwright, C. ; Wainwright, C. ;
Mishkin, P. arXiv preprint arXiv:2307. ; Zhang, Y. ; Zhang, C. ; Agarwal, S. ; Agarwal, S. ; Slama, K. ; Slama, K. Training language models to follow instructions with
human feedback.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Perez, F.; and Ribeiro, I. p. 1 (2023) introduce indirect prompt injection, in which an attacker plants instructions inside data
later retrieved by an LLM-integrated application. Perez and
Ribeiro (2022) demonstrate the underlying ignore previous
instructions attack, and Liu et al. ; and Ribeiro, I. Ignore previous prompt:
Attack techniques for language models.Where the paper says it
Where the paper says it
Perkins, M. p. 0 2021)
has made it trivial for a student, whether in secondary school
or at university, to obtain a fluent answer to almost any shortanswer or short-essay assignment by pasting the prompt into
a chatbot. Surveys in education and rapid reviews of the resulting literature (Cotton, Cotton, and Shipway 2024; Susnjak
and McIntosh 2024; Lo 2023; Perkins 2023; Stokel-Walker
Copyright © 2026, Association for the Advancement of Artificial
Intelligence (www. Perkins, M. Academic integrity considerations of AI
large language models in the post-pandemic era: ChatGPT
and beyond.Where the paper says it
Where the paper says it
Rehberger, J. p. 0 Output watermarking is technically attractive but is only
available to parties who control the model, while classroom
teacher or university instructor generally does not. A third option, less explored in the academic literature
but broadly discussed by the prompt-security community
(Rehberger 2024a), is to watermark the input that the model
is asked to consume. Output watermarking is technically attractive but is only
available to parties who control the model, while classroom
teacher or university instructor generally does not. A third option, less explored in the academic literature
but broadly discussed by the prompt-security community
(Rehberger 2024a), is to watermark the input that the model
is asked to consume. Output watermarking is technically attractive but is only
available to parties who control the model, while classroom
teacher or university instructor generally does not. A third option, less explored in the academic literature
but broadly discussed by the prompt-security community
(Rehberger 2024a), is to watermark the input that the model
is asked to consume. Rehberger, J. ASCII Smuggler Tool: Crafting Invisible Text and Decoding Hidden Codes.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Rehberger, J. 2024b. Hidden Prompt Injections with Anthropic Claude. p. 9 Rehberger, J. Hidden Prompt Injections with Anthropic Claude.Where the paper says it
Rehberger, J. 2024c. Security ProbLLMs in xAI’s Grok: A Deep Dive. p. 9 Rehberger, J. Security ProbLLMs in xAI’s Grok: A
Deep Dive.Where the paper says it
Sadasivan, V. p. 0 Both empirical (Liang et al. 2023) and theoretical (Sadasivan et al. Sadasivan, V. ; Kumar, A. ; Kumar, A. ; Balasubramanian, S. ; Balasubramanian, S. ; and Feizi, S. ; and Feizi, S. Can AI-generated text be reliably
detected?Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Stokel-Walker, C. p. 0 2021)
has made it trivial for a student, whether in secondary school
or at university, to obtain a fluent answer to almost any shortanswer or short-essay assignment by pasting the prompt into
a chatbot. Surveys in education and rapid reviews of the resulting literature (Cotton, Cotton, and Shipway 2024; Susnjak
and McIntosh 2024; Lo 2023; Perkins 2023; Stokel-Walker
Copyright © 2026, Association for the Advancement of Artificial
Intelligence (www. Stokel-Walker, C. AI bot ChatGPT writes smart essaysshould professors worry?Where the paper says it
Where the paper says it
Susnjak, T. p. 0 2021)
has made it trivial for a student, whether in secondary school
or at university, to obtain a fluent answer to almost any shortanswer or short-essay assignment by pasting the prompt into
a chatbot. Surveys in education and rapid reviews of the resulting literature (Cotton, Cotton, and Shipway 2024; Susnjak
and McIntosh 2024; Lo 2023; Perkins 2023; Stokel-Walker
Copyright © 2026, Association for the Advancement of Artificial
Intelligence (www. 2021)
has made it trivial for a student, whether in secondary school
or at university, to obtain a fluent answer to almost any shortanswer or short-essay assignment by pasting the prompt into
a chatbot. Surveys in education and rapid reviews of the resulting literature (Cotton, Cotton, and Shipway 2024; Susnjak
and McIntosh 2024; Lo 2023; Perkins 2023; Stokel-Walker
Copyright © 2026, Association for the Advancement of Artificial
Intelligence (www. ; and McIntosh, T. ChatGPT: The end of
online exam integrity? A rapid review of the literature. Education sciences, 13(4):
410.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The Unicode Consortium. p. 2 of the Unicode Standard (The Unicode Consortium 2022),
which mirrors printable ASCII into a range that virtually
no font renders. The prompt-security community has documented since 2024 that several frontier LLMs tokenize these
tag characters as ordinary text and follow instructions encoded in them, a technique commonly called the ASCII Smuggler (Rehberger 2024a). The Unicode Consortium. The Unicode Standard, Version 15.Where the paper says it
Where the paper says it
Toulas, B. 2025. Google Won’t Fix New ASCII Smuggling Attack in Gemini. p. 9 Mountain View, CA: The Unicode Consortium. Google Won’t Fix New ASCII Smuggling
Attack in Gemini.Where the paper says it
Ueoka, H.; Murawaki, Y. p. 8 Second, combining the Unicode-Tag channel with a contentlevel watermark, in the spirit of (Yang et al. 2022; Ueoka,
Murawaki, and Kurohashi 2021), would survive a student
who lightly paraphrases the brief before pasting it. Second, combining the Unicode-Tag channel with a contentlevel watermark, in the spirit of (Yang et al. 2022; Ueoka,
Murawaki, and Kurohashi 2021), would survive a student
who lightly paraphrases the brief before pasting it. ; and Kurohashi, S. Frustratingly easy edit-based linguistic steganography with a masked
language model.Where the paper says it
Where the paper says it
Where the paper says it
Yang, X.; Zhang, J. p. 8 arXiv preprint arXiv:2307. ; Zhang, Y. arXiv preprint arXiv:2307. ; Zhang, Y. ;
and Yu, N. Tracing text provenance via context-aware
lexical substitution.Where the paper says it
Where the paper says it
Where the paper says it
Yeung, K. 2026. Tokenization Attacks on LLMs. p. 9 In Proceedings of the AAAI Conference
on Artificial Intelligence, volume 36, 11613–11621. Tokenization Attacks on LLMs.Where the paper says it
Zhao, X.; Ananth, P. p. 9 com/research/tokenization-attacks-onllms-how-adversaries-exploit-ai-language-processing. ; Ananth, P. ; Ananth, P. ; and Wang, Y. ; and Wang, Y. Provable
robust watermarking for ai-generated text.Where the paper says it
Where the paper says it
Where the paper says it
Zou, A.; Wang, Z.; Carlini, N. p. 9 arXiv preprint
arXiv:2306. ; Carlini, N. ; Carlini, N. ; Kolter, J. ; Kolter, J. ; and
Fredrikson, M. ; and
Fredrikson, M. Universal and transferable adversarial attacks on aligned language models.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
inputside watermarking for assignment prompts allows an educator to embed an invisible instruction inside a visible assignment prompt, which an LLM will read and write a tell-tale signature into its reply, exposing the copy-and-paste pathway specifically.
-
SteganoPrompt is a single-page zero-dependency web tool that encodes an arbitrary printable ASCII payload into the deprecated Unicode Tags block U+E0000–U+E007F and concatenates it with the visible assignment text. This ensures the encoded string is visually identical to the original on every text renderer, survives common copy-paste channels, and is tokenized as ordinary text by most frontier models.
-
The system allows an instructor to verify the embedded payload before distribution using a Verify panel that computes verify(X) = D rhoT (X), returning W exactly if X is a well-formed watermarked document, thus confirming the literal token presence in a suspect submission.
Abstract
Large language models (LLMs) have made fluent essay writing, code drafting, and quiz answering instantly available to students at every level, from secondary school through graduate study. Many educators do not object to LLM use per se; what they need to detect is the case in which a student pastes the assignment prompt into a chatbot and submits the model's reply verbatim, without engaging with the work. Existing post-hoc AI-text detectors remain unreliable and have been shown to penalise non-native English writers, while output-side watermarks require cooperation from the model provider. We propose an alternative that the educator controls directly: an input-side watermark in which an invisible instruction is embedded inside the visible assignment prompt itself. An LLM that ingests the prompt verbatim quietly reads the hidden instruction and writes a tell-tale signature into its reply, exposing the copy-and-paste pathway specifically. We describe SteganoPrompt, a single-page, zero-dependency web tool that encodes an arbitrary printable-ASCII payload into the deprecated Unicode Tags block (U+E0000--U+E007F). The encoded string is visually identical to the original, survives common copy-paste channels (Word, Google Docs, PDF, Markdown, Slack, e-mail, the major learning-management systems), and is tokenized as ordinary text by most frontier models. We evaluate compliance across eight LLM families and a representative set of educational content channels. To keep deployment transparent, disclosure is built into the tool itself: every encoded brief also carries a brief, always-on academic-integrity reminder for the student, and we set out concrete guidance recommending for its disclosed, ethical use in the classroom. The tool is released under the MIT licence at https://ezharjan.github.io/SteganoPrompt/.
Sources
- On the Opportunities and Risks of Foundation Models
- Defending Against Indirect Prompt Injection Attacks With Spotlighting
- Robust Distortion-free Watermarks for Language Models
- Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews
- Ignore Previous Prompt: Attack Techniques For Language Models
- Can AI-Generated Text be Reliably Detected?
- Provable Robust Watermarking for AI-Generated Text
- Universal and Transferable Adversarial Attacks on Aligned Language Models
Related papers
- From Strategy Search to Policy Learning: MCTS-Guided Adversarial Co-Evolution for LLM Safety Alignment
- Privacy Leakage from Recurrent State in LLM Inference
- Incrementally Verifiable Oracle Aided Computation
- Privilege and confidentiality in generative AI workflows
- BadMoE: Backdooring Mixture-of-Experts LLMs via Routing-Aware Triggers and Task-Underutilized Experts
- CoVault: Secure, Scalable Analytics of Personal Data