Accuracy
Since some AI platforms are not yet connected to the internet, information can be inaccurate, out of date or missing. It's important to check the accuracy of any factual content produced by AI. Pick a few factual statements made in the content and verify them, even if sources were provided. Errors can still slip through.
E.g.: (Prompt asked 14/02/2024)
Authority/Bias
AI models can reflect and amplify societal biases as well as make factual mistakes. When presenting an argument or position using AI-generated content as support, you should carefully evaluate the credibility and diversity of the system's training data, examine who created it and for what purposes, check outputs against other reliable sources, and watch for statements that promote stereotypes or lack nuance.
Oversimplification
AI systems tend to break down complex issues into simpler components to process and respond, but this can miss the nuances, interconnectedness, or wider context around topics. The data used to train AI systems itself often lacks nuance or depth. So the responses the AI generates tend to reflect the simplified nature of the data rather than the full complexity of the real world. Ensure you are reading more on any topic that AI output describes to ensure you have a full grasp on the concept/topic.
Plagiarism
Many AIs summarise content they find online, which could plagiarise. Students should keep a critical eye on whether AI-generated text credits its sources or simply regurgitates and recombines content from across the internet. To evaluate, students can search snippets of text to check if passages appear word-for-word elsewhere online.
Check for American English
Many AIs are currently trained to produce American English responses, therefore before you use any text that you plan to share / be used by others, ensure the spelling is correct for British English.
As you utilise AI tools to assist with your academic work, it is important to critically evaluate the content these tools produce, just as you would any other source.
The skills required to effectively assess AI-generated text and data are similar to those you already employ when determining the credibility and relevance of online sources for your research papers and assignments.
Evaluating the strengths and limitations of AI writing and data aids will allow you to use these emerging technologies appropriately as one part of a larger academic inquiry process.