ARTICLE TEMPLATE Beyond Detection: Designing AI-Resilient Assessments with Automated Feedback Tool to Foster Critical Thinking and Originality Muhammad Sajjad Akbar a a University of Sydney, Australia; ARTICLE HISTORY Compiled April 1, 2025 ABSTRACT The growing prevalence of generative AI tools such as ChatGPT has raised urgent concerns about their impact on student learning, particularly their potential to erode critical thinking and creativity in academic contexts. As students increasingly use these tools to complete assessments, foundational cognitive skills are at risk of being bypassed, challenging the integrity of higher education and the authenticity of student work. Current AI-generated text detection tools are fundamentally inadequate in addressing this challenge. They produce unreliable, unverifiable outputs and are highly susceptible to false positives and false negatives, especially when students apply obfuscation techniques such as paraphrasing, translation, or structural rewording. These tools rely on shallow statistical features rather than contextual or semantic understanding, making them unsuitable as definitive indicators of AI misuse. In response, this research proposes an AI-resilient, assessment-based solution that shifts focus from reactive detection to proactive assessment design. The solution is delivered through a web-based Python tool that integrates Bloom's Taxonomy with advanced natural language processing techniques including GPT-3.5 Turbo, BERT-based semantic similarity, and TF-IDF metrics to evaluate the AI-solvability of assignment tasks. By analyzing both surface-level and semantic features, the tool helps educators assess whether a task targets lower-order thinking (e.g., recall, summarization), which is more easily completed by AI, or higher-order skills (e.g., analysis, evaluation, creation), which are more resistant to AI automation. This framework empowers educators to intentionally design cognitively demanding AI-resistant assessments that promote originality, critical thinking, and fairness. By addressing the design of root issue assessment rather than relying on flawed detection tools, this research contributes a sustainable and pedagogically sound strategy to uphold academic standards and foster authentic learning in the era of AI. KEYWORDS Generative AI; ChatGPT; AI-resilient; Bloom's Taxonomy; Automated Assessments; AI-solvability;Automated Feedback; appendices 1. Introduction Integrating AI-technology with innovative thinking skills in higher education (HE) environment has grown more challenging due to rapid digital innovation and ubiquitous data availability. In applied education, innovative thinking is essential. It is charac-CONTACT Muhammad Sajjad Akbar. It entails thinking creatively to come up with original solutions to issues, enhance workflows, or open up new possibilities.

AI is the big new thing, and already, folks are coming up with countless ways they can use it. Some of these ideas, though, are downright ludicrous. While AI-powered text generation tools like ChatGPT, Copy.ai, and Jasper are incredibly impressive, their usefulness and practicality are very easy to oversell. I've seen countless "suggestions" in viral blog posts and Twitter threads that don't work--but look like they did because the generated text seems plausible. So, if you're thinking about adding any AI content generators to your workflow, there are a few things you should keep in mind.

AI text generators, also known as language models, are algorithms that use artificial intelligence to generate human-like text based on a given prompt or seed text. These models are trained on vast amounts of text data, learning patterns, and relationships within the data to produce coherent and meaningful responses. AI or ML text generators can be used in a wide range of applications, including chatbots, customer service, content creation, and even creative writing. The Global AI Text Generator Market was valued at USD 360 million in 2022 and is expected to grow at a CAGR of 18% during the forecast period of 2023-2032 to reach USD 1,808 million. AI/ML generators can be used to produce responses to user queries in a conversational and natural manner, making them useful in developing chatbots for customer service or online support.

AI-generated text, from tools like ChatGPT, is starting to impact daily life. Teachers are testing it out as part of classroom lessons. Marketers are champing at the bit to replace their interns. Memers are going buck wild. It would be a lie to say I'm not a little anxious about the robots coming for my writing gig.

Are you looking to improve the quality of your written content, but don't have enough time or resources to make it happen? If so, you might want to consider using an AI text generator like Jasper or WordHero. Not only are they a quick and easy solution for turning out engaging content in record time, they can also help ensure that your website is optimized with SEO-friendly keywords. In this blog post, I will examine the differences of Jasper vs WordHero by taking an in-depth look at how these two AI text and content creation tools compare against one another when it comes to features, pricing structure, ease-of-use and more. AIgeeked.com is reader supported and we may earn an affiliate commission from the AI products listed on this website. If you're a content writer, blogger, or SEO specialist in need of an AI writing tool to help you create high-quality, SEO-friendly content, chances are you've been debating between two choices: Jasper and WordHero.

You can see where this is headed. A writing assignment asks students to compare and contrast feminist themes in Jane Eyre and Wuthering Heights. Yup, it can do that. A political science exam requires short-essay responses to questions around the rise and fall of the Soviet Union. So what does this all mean for teaching?

September 2022 was apparently the month artificial intelligence essay angst boiled over in academia, as various media outlets published opinion pieces lamenting the rise of AI writing systems that will ruin student writing and pave the way toward unprecedented levels of academic misconduct. Then, on Sept. 23, academic Twitter exploded into a bit of a panic on this topic. The firestorm was prompted by a post to the OpenAI subreddit where user Urdadgirl69 claimed to be getting straight A's with essays "written" using artificial intelligence. Professors on Reddit and Twitter alike expressed frustration and concern about how best to address the threat of AI essays. One of the most poignant and widely retweeted laments came from Redditor ahumanlikeyou, who wrote, "Grading something an AI wrote is an incredibly depressing waste of my life."

Writing already involves AI, in the form of predictive text for example. We are familiar with this on our phones and in our emails. Humans have been collaborating with technology for writing since sticks were used for drawing in sand or on cave walls. Ingenuity and creativity mean that this technology is constantly changing. Most recently, the advent of AI writers (software that uses artificial intelligence to generate text) has created a lot of excitement and concern.

OpenAI, the non-profit founded by Elon Musk in 2015 – he's no longer part of it – has released the biggest and final version of the GPT-2 text-generating language model, which it has admitted could be dangerous in the wrong hands. However, it says the newly released full model's output is only slightly more convincing to humans than the previous version. The organization released the first portion of the model in February as part of a staged process, beginning with just 124 million parameters. It held back the full model with 1.5 billion parameters because scientists believed it was too dangerous and could be used by malicious actors, such as terrorists and state-sponsored hackers. Among the malicious purposes for which OpenAI admitted GPT-2 might be used are generating misleading news articles, impersonating others online, automating the production of abusive or fake content for social media, and automating the creation of spam and phishing content.

There's no denying it at this point: AI is going to replicate media (videos, articles, pictures, etc.) with such authenticity that unaided people will not be able to tell the difference between what's real and what's not. We've already seen multiple examples of insanely realistic deepfake videos, and now it appears as if we have the text equivalent of those fakes with this AI-powered text generator, dubbed Talk to Transformer. Before reading on, open up that link to Talk to Transformer in a new tab and pop in your own prompt. The site, which was created by machine learning engineer Adam King, will take somewhere between five to ten seconds to load, and then, from just the half-sentence you've written, will spit out a couple hundred words (rough average number of words for our test prompts) that are astoundingly coherent. Talk to Transformer is able to generate such humanlike text thanks to--you probably guessed it--neural networks coupled with big data.