How Does Blackboard Detect AI?

Can Blackboard detect AI writing?

Not on its own, but it has the tools to catch AI text. Blackboard is just a platform for uploading, downloading, and sharing educational materials. It’s not designed to actually inspect anything on its own. However, teachers can (and almost universally do) install programs in their Blackboard classrooms to catch AI writing and plagiarism. So, while Blackboard can’t technically detect AI on its own, the tools to catch AI are absolutely available on the platform. The two biggest and most universally accepted cheating detectors are SafeAssign, and TurnItIn. These days, most educators use one (or both) of these tools.

Why do plagiarism detectors catch AI?

Anything AI shares with you is borrowed from somewhere else. AI tools like ChatGPT, Bard, Claude, Bing AI, and Gemini aren’t writing anything from scratch. They take your question, scan the internet and their archives to find information, and then repackage that info in a conversational answer. It’s cool, but it’s pulling info from things that are already written. That means that any material you copy and paste from an AI chatbot is potentially already published somewhere. Even if you reverse search a passage or text to confirm that it isn’t already published somewhere, the text may still be flagged as plagiarism.

AI repeats itself often, so even unborrowed answers may be published. Let’s say you know for an absolute fact that the content you plan on taking isn’t published anywhere. It may still get flagged as plagiarism! AI tools won’t give unique answers for the same question, so if anyone else has ever asked a chatbot the question you just asked, the material could already be in a plagiarism detector’s archives. In other words, if you asked AI, “How did tax policy influence the American Revolution” and then copy and paste the answer, it’s possible that some other student asked the chatbot the same question months ago! If they also had to submit the paper to Blackboard, your work will be flagged.

How does SafeAssign catch AI writing?

SafeAssign has an algorithm that compares texts for similarities. SafeAssign also has a database of more than 59 million student essays, so it can even tell when someone is stealing text from something that has never been published. Beyond that, papers submitted to SafeAssign will be tested against all published materials online. It even checks for minor variations in a piece of writing, so borrowing text from AI and changing a few words won’t trick it. SafeAssign doesn’t give teachers a “Hey, this paper is plagiarized” indicator. It marks passages of a text with a percentage (i.e. “71% match with another text”) and then the teacher can look at the source material to see if it has been stolen or not. So, plagiarizing even a short passage will get you flagged.

How does TurnItIn catch AI writing?

TurnItIn analyzes writing style and tone to detect AI. While SafeAssign compares texts, TurnItIn analyzes the content of an essay and works backwards. Every AI chatbot has its own nuances, idiosyncrasies, and style. If TurnItIn scans a sentence and it seems like a particular chatbot might have written it, it will scan the rest of the text to see if there’s a pattern. Then, it will flag the text in question for review. TurnItIn is 97% accurate when it comes to identifying AI writing. It only has a false positive rate of 1%, so if you use any AI writing at all, it’s extremely likely that you’re going to be caught. TurnItIn will also run a normal scan for plagiarism the same way that SafeAssign does.

Can you get around AI detection?

The available free tools for bypassing AI are very risky. There are a handful of free websites that claim they’re able to trick AI detectors, but the quality of the writing that these websites churn out is pretty poor. On top of that, it’s not entirely clear that they’re actually able to trick any AI detectors. Tip for teachers: A lot of these free websites just shift sentences from the active to passive voice, so if you see a student’s work randomly shifting voice out of nowhere it’s possible the student is trying to hide AI text. Tip for teachers: Look for odd word choices with homophones (i.e. “The subject of this essay” becomes “To subject you to this essay.”). AI bypassing tools tend to mess up intended word usage, so if you spot strange mistakes with word meanings, it could be an AI bypassing tool.

Paid rewriting programs are expensive and questionable. The most “proven” way to get around AI detectors is to use a paid program built for such a purpose. Programs like Bypass AI or Bypass GPT are more efficient than free options, but they are expensive (usually around $30 a month on the low end). They’re also not 100% likely to beat AI detectors as comprehensive and powerful as TurnItIn and SafeAssign, so it’s still not a great option.

Misspelling words and adding grammar mistakes can help. AI tools may not be very good at writing elegant essays, but it doesn’t tend to misspell or make grammar mistakes. If you are completely dedicated to tricking an AI detector, just put a typo in every other word. Tip for teachers: Make grammar and spelling a part of the scoring rubric for every essay you assign. This should prevent major issues with AI in the classroom.

Is it ethical to use AI for schoolwork?

Absolutely not, and you’re only doing yourself a disservice by using it. Using AI to complete your schoolwork is plagiarism, and it’s wrong to turn in work as your own when it’s not. But beyond that, you’re only robbing yourself of an education if you skip assignments by cheating. Even the most frustrating essay assignment is helping you cultivate the critical thinking and writing skills you’ll need to achieve your dreams, so don’t cheat!