Home › Resources › Does Blackboard Detect AI?

Does Blackboard Detect AI?

Short answer: no. Blackboard is the one major learning platform that ships its own originality checker, SafeAssign, and SafeAssign is not an AI detector. That is not a gap Anthology is racing to close. Anthology tested AI detection, decided against building it into Blackboard, and published its reasoning, which makes Blackboard the only LMS whose vendor is on record against the tools students are most worried about.

Check your own writing free How our detector works

That single fact resolves most of the confusion. Because Blackboard genuinely does have a built-in checker, people reasonably assume the checker covers AI. It does not, and it structurally cannot. Understanding the difference between a similarity check and an AI check is the whole answer, and it is also the thing most worth knowing if you ever have to discuss a flagged assignment.

Below: what SafeAssign really compares your work against, why a plagiarism checker cannot catch ChatGPT even in principle, Anthology's own published position on AI detectors, and where an AI score in Blackboard can come from if you do see one.

Illustration of a student holding up two printed papers side by side, comparing them, with document and globe icons above. The caption reads: does Blackboard detect AI? No. It checks copying. — SafeAssign looks for text that already exists somewhere. That is a different question from who wrote it.

The short version

Two different questions get asked with the same words.

Does Blackboard check for plagiarism? Yes. SafeAssign is built in, at no extra cost, and many institutions have it on by default.
Does Blackboard check for AI writing? No. SafeAssign produces a matching score against existing sources. It does not produce an AI score, and Anthology chose not to add one.

Anthology's documentation describes SafeAssign as a tool that "compares submitted assignments against a set of academic papers to identify areas of overlap between the submitted assignment and existing works." Overlap with existing works is the entire job. Nothing in that sentence is about authorship.

So if you have seen a SafeAssign percentage and wondered whether it was an AI verdict: it was not. It was a measure of how much of your text matches text that already existed somewhere.

What SafeAssign checks, in one paragraph

SafeAssign compares a submission against four collections: the open internet, the licensed ProQuest ABI/Inform scholarly database, your own institution's archive of previously submitted work, and the cross-institutional Global Reference Database. It returns an Originality Report with an overall matching percentage and a per-source breakdown.

A match is not automatically plagiarism. Correctly quoted and properly cited material matches too, as do reference lists, standard phrases and an assignment prompt you were told to restate. The per-source list matters far more than the headline number.

We cover SafeAssign itself in depth on a separate page: what each of the four collections holds, how the Global Reference Database opt-in works, how SafeAssign differs from Turnitin, and how to read a matching percentage. See the SafeAssign AI checker explained.

The rest of this page is about Blackboard the platform: Anthology's published decision on AI detection, the third-party tools an institution can bolt on, and the proctoring layer that gets confused with all of it.

Why a similarity checker cannot catch ChatGPT, even in principle

This is the part worth being precise about, because it is the reason no amount of SafeAssign improvement would change the answer.

A similarity checker works by matching. It finds strings of your text that also appear in a source it holds. Its entire capability rests on the copied passage existing somewhere else first.

Text generated by a language model does not exist anywhere else. It is assembled fresh for your prompt. There is no prior document to match it against, so a matching engine has nothing to find. An essay written entirely by ChatGPT can legitimately return a very low SafeAssign score, and that is the system working exactly as designed, not failing.

A low SafeAssign percentage means "this text does not appear in our sources." It does not mean "a human wrote this." Those are different claims, and only the first one is being made.

AI detection works on a completely different principle. Instead of matching, it scores statistical properties of the writing itself, how predictable each word is given the ones before it, how evenly distributed the sentence rhythms are. That is a probability estimate about style, not a lookup, which is why it carries a false positive rate in a way a matching engine does not. We explain the mechanism in how AI detector accuracy actually works, and the failure modes in can AI detectors be wrong.

Illustration about SafeAssign having nothing to match against, showing a student with documents and database icons. — A matching engine can only find text that already exists. Model output is written fresh for your prompt.

Anthology's published position on AI detection

Most LMS vendors have stayed quiet. Anthology, which owns Blackboard, has not, and its position is unusually direct.

Anthology did not just decide against AI detection, it ran a test first and published the numbers. Its white paper, AI, Academic Integrity, and Authentic Assessment, describes a beta test run across May and June 2023 with market-leading AI detection tools, explicitly to evaluate putting AI detection inside SafeAssign. 65 client institutions submitted more than 1,000 texts, some authentic and some AI-generated, then reported how accurately each was assessed.

The headline result: 80 percent of respondents felt the detectors were, at best, only able to "sometimes" identify texts correctly. Anthology records that participants left the test with very low confidence in the ability of AI detectors to tell AI and human writing apart, and that Anthology and its client partners together concluded that AI detection is "not currently fit for purpose in education".

The same paper sets out the outside research it was testing against, which is unusually candid for a vendor document. It cites Sadasivan and colleagues at the University of Maryland, who asked whether AI-generated text can be reliably detected and concluded it cannot, with simple paraphrasing enough to evade detection. It cites Weber-Wulff and colleagues, whose study of 14 detectors across six countries found accuracy ranging from just 33 to 81 percent depending on provider and method. And it cites the Stanford work on bias against non-native English writers that we return to below. Anthology therefore decided not to make AI detection available natively in Blackboard Learn.

What Anthology recommends instead is pedagogical rather than technical: authentic assessment designed around critical thinking and personal reflection, assignments an AI cannot easily answer, peer assessment and group work that adds accountability, and personalised tasks where an instructor already knows a student's voice. Its own AI product, the AI Design Assistant, helps instructors build rubrics, questions and course structure. It generates material for teachers. It does not judge students.

Why this matters if you have been flagged

If you are a student in a Blackboard institution facing an AI accusation, this is a genuinely useful thing to know. The company that sold your institution its learning platform has published that AI detectors are too unreliable and too biased for high-stakes use. That does not settle a case by itself, but it is a documented vendor position rather than a student's opinion, and it belongs in a calm written response. Our AI detection appeal letter guide shows how to structure one, and being accused of using AI in college covers the first 48 hours.

Illustration of an educator raising a hand to say stop, beside a crossed-out robot icon and a set of scales. The caption reads: Blackboard's owner tested AI detectors. Said no. — 65 institutions, more than 1,000 texts, and a published conclusion. Tested in May and June 2023.

So where would an AI score in Blackboard come from?

You can still see an AI percentage inside Blackboard. If you do, it did not come from Blackboard or from SafeAssign.

A third-party tool added by your institution. Blackboard supports external tools through the LTI standard, and both Copyleaks and Turnitin market Blackboard integrations. Whether one is wired in is an administrator's decision, made institution by institution, sometimes department by department.
A tool used outside the LMS entirely. An instructor can paste your text into any consumer AI detector in a browser tab, with no institutional configuration, no audit trail and no accountability. This happens, and it is worth naming, because it produces numbers that nobody can later explain or reproduce.

Which means the first question about any AI number in Blackboard is simply: which product generated this, and what does that product publish about its own error rate? If the number came from Turnitin, for example, Turnitin itself reports a document-level false positive rate under 1 percent only for documents assessed at 20 percent or more AI, shows no percentage at all below that threshold, and reports approximately 4 percent across the full distribution of documents, which we read clause by clause in is Turnitin's AI detector accurate. If the number came from a free web tool, there may be no published rate at all. Those are very different pieces of evidence, and they should not be treated as interchangeable. We compare what the main vendors actually claim in the detector comparison.

Proctoring and lockdown browsers are a separate system

The other source of "Blackboard catches cheating" stories has nothing to do with writing at all.

For timed tests, many institutions bolt a proctoring layer onto Blackboard: a lockdown browser that restricts what you can open during an attempt, webcam monitoring, screen recording, or a combination. Respondus LockDown Browser is a common example. These tools are real, they do flag certain exam behaviours, and they are licensed separately from Blackboard.

They are also completely silent about essays. A lockdown browser restricts a live exam session. It does not read a term paper you uploaded last week and form a view about who wrote it. When someone says Blackboard caught a student cheating, it is worth establishing which of these two systems they mean, because the evidence, the error modes and the appropriate response are entirely different.

Illustration contrasting a locked-down exam window with essay marking, above the caption: LockDown Browser is not an AI detector. — Lockdown guards a live exam session. It never opens a term paper.

How to check what is actually enabled on your assignment

Blackboard comes in two interfaces, Ultra and Original, and the signals look slightly different, but the logic is the same in both.

Look for the SafeAssign indicator on the assignment. When SafeAssign is enabled you will typically see it named on the assignment, and after submission an Originality Report becomes available, often with a short processing delay. In Ultra it surfaces alongside the submission in the grading view; in Original it appears in the Grade Centre.
Check whether you were asked about the Global Reference Database. That prompt only appears when SafeAssign is running and the institution has enabled the option, so it is a reliable tell.
Look for anything named other than SafeAssign. A separate similarity or AI panel, a different vendor logo, or a link that opens a tool outside Blackboard, all indicate a third-party integration rather than the built-in checker.
Read the syllabus and the course integrity statement, then ask. Instructors are usually required to disclose what they use. Asking how your work will be assessed is a normal question, not an admission.

If you want to see how your own writing reads before any of this applies to it, run the draft yourself. TextSight's free AI detector returns a probability with sentence-level highlights, so you can find and rewrite the passages that read as flat rather than guess at them, and the document detector accepts the file directly if your submission is a document. We are explicit about the limits in AI detection limitations, and about how we test in our accuracy methodology. Keeping draft history in Google Docs or Word remains the single strongest authorship evidence you can hold, as how to prove you did not use AI sets out.

How Blackboard compares with Moodle and Canvas

The three platforms have taken three different routes, and Blackboard's is the most opinionated of the three.

Blackboard ships a similarity checker, refuses to ship an AI detector, and has published why.
Moodle ships neither. Its documentation states it comes with no pre-installed plagiarism prevention at all, so everything depends on which plugin an administrator installed. See does Moodle detect AI.
Canvas ships no checker but does ship the integration layer, a plagiarism framework that lets external tools post originality reports back into the gradebook. See does Canvas detect AI.

For educators weighing this up, our guidance for teachers and Turnitin compared with TextSight go into how to use a score without over-trusting it. For students, the student guide and AI detection at college level are the better starting points.