DAU updated Directive 704, Student Academic and Administrative Policies and Procedures Handbook, to include paragraph 6 Academic Integrity (b) (1) The use of Artificial Intelligence (AI) generated material. In the digital age, the use of AI to generate material - from text, pictures, songs, and even algorithms - has become both a tool and a challenge. From academic papers to business reports, AI-generated content is pervasive, and its presence raises important questions about authenticity and credibility. Recognizing AI-generated responses and documents is crucial for maintaining integrity in various fields. But how can you tell the difference between AI and human generated material? By knowing what to look for! This blog outlines practical ways to identify such content. It is not all encompassing as the identification process will evolve as AI continues to change.

I started playing with AI in late 2022 when my son explained how exciting this new technology was going to be for the future. Being a techy geek (not an expert by any stretch), I had to test it out to see what all the rave was about. I immediately could see so many uses for the small options that were available. Since then, AI has expanded well past simple text to AI generated pictures (like the one for this blog), songs, music, excel formula generators and a plethora of other tools that are available – some free; some at an expense to the user. So, how do you know if your learners are using AI generated text? 

To identify AI-generated text, one must first understand the characteristics of AI-generated text. AI-generated text, though increasingly sophisticated, often exhibits certain characteristics that can set it apart from human-generated content. These include a lack of deep understanding, repetitive phrasing, and sometimes an overuse of common phrases. AI text generators typically rely on patterns in data they have been trained on, rather than genuine comprehension of the subject matter. 

One of the most telling signs of AI-generated text is inconsistency in style or voice. AI may struggle with maintaining a consistent narrative voice or may shift tones inappropriately within a single document. Humans, on the other hand, typically have a distinctive and consistent style influenced by their unique perspectives and experiences.

AI-generated responses might lack depth or show a superficial understanding of complex topics. They can fail to capture nuanced arguments or provide in-depth analysis based on critical thinking. For instance, AI is not suited for critical thinking exercises or assignments. Moreover, AI might introduce factual inaccuracies or contextually inappropriate references, which do not align with the known facts or the typical discourse surrounding a topic.

Human writing often reflects emotional intelligence and an ability to connect with the reader on a personal level. AI-generated text, while it can mimic some forms of emotional expression, usually lacks genuine emotional depth or a truly empathetic tone. The text might seem technically competent but emotionally flat.

As AI text generated becomes more advanced, so too do the tools designed to detect it. Several online platforms now offer services to help identify AI-generated text. These tools analyze the writing for patterns typical of AI, such as certain syntactic structures or the predictability of word choices. However, while helpful, these tools are not foolproof and should be used in conjunction with other methods. These tools may claim ownership or authoring credits for works written and published by human authors. 

AI might produce content that seems plausible but falls apart under scrutiny. Cross-referencing information and fact-checking against reputable sources can uncover inaccuracies that suggest a piece of content was AI-generated. This method is particularly effective in academic and technical fields, where precision and accuracy are paramount.

When in doubt, consulting with experts in the relevant field can help determine whether content was likely generated by AI. In many instances, the DAU professor is the expert. It is incumbent upon us to understand and experience AI to discern whether or not our learners are depending on AI generated text to complete assignments. Experts can assess the depth of knowledge, the appropriateness of the methodologies discussed, and the accuracy of the content, providing a human perspective that is currently beyond the capability of AI.

As AI continues to evolve, distinguishing between human and AI-generated content will become increasingly challenging but no less important. By paying attention to the depth of content, consistency of style, emotional engagement, and by using available tools and expert advice, one can become more adept at identifying AI-generated responses and documents. This vigilance is essential not only for academic integrity but also for maintaining trust and authenticity in all fields of communication.

Skills you need to work with AI

Executive Order: President Biden issues EO on safe, secure, and trustworthy AI.