How Do Professors Check for AI?

Artificial intelligence (AI) has certainly been making its mark in education, with 46.9% of college students using the technology in their studies. 

However, not all professors are on board with using AI for coursework, especially when it's used to generate entire assignments.

To help address this issue, professors have started looking for AI writing in coursework. 

So, how do professors check for AI? Well, that depends on a variety of factors, including educational institution policies for AI detection and preferences.

Read on to discover why professors check for AI in assignments and the different strategies they may use to detect it.

Key Takeaways (TL;DR)

• Professors check for AI in assignments for various reasons.
  • Preventing cheating.
  • Avoiding unfair advantages in assignment completion.
  • Properly assessing a student’s understanding of course material.
• How professors check for AI may include a variety of methods (depending on college and university policies around AI detection), including:
  • Looking for inaccurate or fake references and sources in assignments.
  • Significant changes in student writing style from previous work.
  • Professors may generate their own AI-written assignments for comparison. 
  • Discuss with students how and if AI was used, such as editing. 
  • Use AI content detectors to determine the likelihood of AI use in assignments.

Why Do Professors Check for AI in Assignments?

There are three main reasons why professors check for AI in assignments:

• To detect and discourage cheating. With the right prompt, generative AI tools like ChatGPT can write entire essays. Students may then try to pass off these AI-generated essays as their original work, which many see as cheating.
• To level the playing field. Students with access to advanced AI tools may use them to create more high-quality assignments than those without.
• To accurately assess a student’s abilities. It’s hard for a professor to assess a student’s understanding of the material if a chatbot is providing the answers for them.

How Do Professors Check for AI?

AI content is difficult for humans to identify, so professors may use a combination of strategies to check for AI in student work.

Check references

A good, well-researched paper often has plenty of references to support its claims. However, when ChatGPT cites its sources, they aren’t always accurate.

One study found that out of 115 references generated by ChatGPT, just 7% were real and accurate. Because of AI hallucinations like this, professors may verify the authenticity of sources in a student’s reference list if they suspect an AI-generated paper.

Make comparisons to previous work…

You won’t often see two papers written exactly the same way. Each student has a unique writing style, and professors may use this to help them identify AI-generated content.

However, there are a few additional considerations with this approach.

• If a professor is unfamiliar with a particular student’s work, they may not recognize changes in writing style (such as that beginning of a term).
• Large class sizes may mean that multiple educators and educational assistants are evaluating papers and don’t become as familiar with each student’s writing style.
• If instructors notice a complete or sudden shift in style, it doesn’t necessarily mean that AI was used. The student could have a writing style that’s simply inconsistent.

…and AI-generated text

Using generative AI tools isn’t specific to students — professors are also often familiar with it.

They could use ChatGPT to create one or even multiple versions of an essay based on their specific essay guidelines and then compare the outputs to student assignments.

However, as with comparing writing styles, there are issues with this method. AI tools don’t often produce the same output when provided with the same prompt. So, it’s unlikely to provide an obvious example of AI use.

Ask questions

Sometimes, professors check for AI by having a conversation with students and asking questions such as:

• How did you plan the assignment? Can you share rough notes, outlines, or drafts?
• Where did you get the ideas for your thesis?
• What sources did you use? Can you share how you researched them? 
• Did you use AI when working on the assignment, such as for editing? Can you share an example?

Being able to provide rough notes, drafts, brainstorming, a list of sources, and clarity on any AI tools incorporated, such as for editing grammar or spelling, can help to communicate the authenticity of work.

Additionally, students may demonstrate the writing process from beginning to end if they write in Google Docs and use the Originality.ai Chrome Extension.

Use AI content detectors

Only about 25% of teachers feel very effective at distinguishing between AI and student-written assignments. 

Though AI detection policies for assignments can vary by college or university, where it’s part of policies, some professors may use an AI content detector as well.

The technology behind how an AI detector works is complex, with advanced algorithms and machine learning playing key roles. However, using an AI detector like Originality.ai to scan for AI content is easy and intuitive.

According to several third-party academic and research studies, Originality.ai demonstrates exceptional accuracy in AI detection. However, false positives are possible, so an AI detector shouldn’t be the only method for establishing academic standing.

Learn more about Originality.ai for educators.

Final Thoughts

As AI content continues to rise, professors, educators, and students must establish clear communication around AI in assignments and work together to navigate it properly. AI detectors for educators are a key way to improve transparency about AI content in the classroom.

Additionally, tools like the Originality.ai Chrome Extension are a fantastic way to demonstrate the authenticity of work by visualizing the writing process from beginning to end.