Packback's AI Risk indicator is a tool designed to help you assess the likelihood that a student's text was generated by artificial intelligence (AI). This indicator analyzes the student's writing style, looking for patterns that are characteristic of AI-generated text.

How it Works:

Packback's algorithm examines the student's word choice and sentence structure. AI language models tend to follow predictable patterns, while human writing is often more nuanced and varied.

Interpreting the AI Risk Indicator:

• Low: A low risk level suggests that the submission is likely human-generated. However, it's important to remember that AI-generated text can sometimes be difficult to detect.
• Medium: A medium risk level indicates that the algorithm has identified some patterns that could be indicative of AI-generated text. This warrants a closer look at the student's work. If you suspect AI usage, consider having a conversation with the student about their writing process.
• High: A high risk level means that the algorithm has detected strong evidence of AI-generated text. This doesn't necessarily mean the student intentionally used AI to complete the assignment, but it's important to investigate further.

Important Considerations:

• False Positives: It's possible for human writing to be flagged as moderate or even high risk. This is more likely to occur with writing that is very concise, formulaic, or uses repetitive language. Packback intentionally tunes our model to reduce the likeness of false positives,
• Evolving Technology: AI writing tools are constantly improving, making it more challenging to detect AI-generated text. Packback's algorithm is continually updated to stay ahead of these advancements.

What happens when my students use the feedback from Packback’s AI writing assistant in Deep Dives or Writing Lab?

Packback's AI writing assistant is designed to provide feedback and suggestions on grammar, structure, and clarity. It helps students improve their own writing, not replace it. Unlike some AI tools that generate entire paragraphs or essays, Packback's assistant doesn't write original content. It simply helps students refine their existing work. Any modifications students make based on the feedback are still ultimately their own. They're the ones making the decisions about how to revise and improve their writing. Therefore, the use of Packback’s AI writing feedback in the platform should not impact students’ AI Risk indicator. 

Using the AI Risk Smart Highlighter:

When an assignment has an overall AI Risk of Medium or High, a smart highlighter panel will appear in the AI Risk Report to show which parts of the text are most likely to contain AI-generated text. Educators can click the arrows in the smart highlighter to view each section of text that potentially contains AI.

Please Note: Currently, the highlighter is only available for educators when evaluating assignments. In the future, it will be available to students as well.

Using the AI Risk Indicator Effectively:

The AI Risk indicator is not intended to be a definitive judgment of whether a student used generative-AI. Instead, it's a starting point for further investigation and conversation.

• Review the student's work: Carefully read the submission, paying attention to the writing style, originality of ideas, and overall quality.
• Talk to the student: If you have concerns about AI usage, have an open and honest conversation with the student. Encourage them to discuss their writing process and any resources they used.
• Use your judgment: Combine the AI Risk indicator with your own knowledge of the student and the assignment to make an informed decision about whether AI was used.

Remember: The goal is not to punish students for using AI, but to encourage them to develop their own writing skills and engage in authentic learning. By using the AI Risk indicator thoughtfully, you can foster a classroom environment that values both technological advancement and academic integrity.

 

Disabling AI Risk Detection

Educators can choose to disable the AI Risk indicator and report when creating an assignment. Uncheck the box next to “Enable AI Risk report in the Originality Fingerprint” to disable the feature. When disabled, students and educators will not see the AI Risk indicator or report.

PLEASE NOTE

Packback recently upgraded our AI writing detection model to increase its efficacy. If you’re viewing a Writing Assignment submission that already had an AI Risk report generated using the previous model, you can regenerate the report to view the new one. See this article for more details.