Surface-Level Synthesis, Lack of Critical Depth, and Errors: Why AI-Based Analysis May Be a Bad Idea
The plan was good. You use a well-designed prompt found online. You feed it to the best AI platform suggested on student forums. In just several minutes, you get a well-structured literature review, essay or even a PhD section draft. You double-check the outputs for fake references and other well-known problems. Everything looks neat, well-structured, and professional. However, your supervisor does not look happy. They use words such as ‘shallow’, ‘derivative’, ‘superficial’ or ‘lacking criticism’ to describe your results. You also feel that something about your work feels ‘off’ while you can’t put your finger on the cause with absolute certainty.
In this article, we will analyse the main reasons contributing to such sub-optimal results. These facts are not instantly obvious to non-professional AI users. Yet, there are some valid reasons why AI-based analysis may be a bad idea for your academic work.
The Quality of Synthesis
Imagine a new worker on a probationary period who really needs a certain job but is not sufficiently qualified for the position. To stay employed, they demonstrate the highest levels of motivation and diligence possible. The problem is that these qualities alone are not enough to produce superior results. The same is true for AI. Despite your carefully designed prompts, it still lacks the expertise and insight required to develop a strong, independent argument.
As a result, the synthesis of data produced by AI only grasps surface-level ideas. This may be sufficient for developing an initial understanding of certain areas. However, this is not sufficient for producing a well-rounded synthesis of sources and linking it with your own academic novelty. If you are not aware of these limitations, you can easily end up with a literature analysis going nowhere in terms of setting the scene for your original thought.
The Lack of Criticism
One of the premises behind AI platforms development is giving the user a coherent and accurate answer to a certain question. If you ask the algorithm what the height of the Eiffel Tower is, the algorithm will be happy to provide a clear and unambiguous response. Problems begin when there are no right and wrong answers. In academic research, especailly PhD level writing, you frequently encounter multiple theories and standpoints with contrasting views. Unfortunately, AI systems are not notoriously good at handling paradoxes or ambiguity in general.
This leads to the overall lack of criticism in AI synthesis. Some ideas that appear statistically more prominent usually have a greater chance of overrepresentation. Additionally, AI systems do not appraise the quality of the sources they can access. In this aspect, a Q1 article from a reputable peer-reviewed journal may be deemed equal to an industry publication or a blog post. With such estimations frequently occurring within a ‘black box’, your control over the outputs may not be as absolute as you expect.
Errors
Every researcher attempting to use AI for critical tasks has encountered the problem of fake references or incorrect article summaries. They are difficult to notice and require substantial time to identify. With that being said, interpretive and logical errors may be as dangerous as factual ones. If an AI does not understand the complex meaning of a certain term, it either omits it or interprets it incorrectly. Such flawed premises lead to compounding errors permeating all layers of your draft, from source review to analysis results.
If your knowledge of a certain field is limited, you are not equipped with the expertise required for identifying such interpretive and logical errors. Moreover, you will not be able to answer questions from your supervisor knowledgeable in this field who is well versed in the key authors and theories in it. In some cases, you may even have to seek the services of a highly skilled PhD academic writer to sort out interpretive errors and fix a deeply flawed draft you have limited control over.
Losing Your Own Grasp
Last but not least, the use of AI quickly kills your own ‘analytical muscle’. You can think of this as an energy drink or a cup of strong coffee. Both of these beverages are great for carrying you through the exhausting day of preparing for an important exam or a one-day jet-lagged business trip. Yet, you would not want to consume extreme amounts of energy drinks or drink dozens of espressos regularly. Unfortunately, this is exactly what many students do before they realise the detrimental effect of AI on their writing skills.
According to multiple studies, artificial intelligence is increasingly used as a ‘cognitive offloading’ tool rather than a mechanism facilitating repetitive or mechanical tasks. From an academic writing standpoint, this creates dependence on such instruments and reduces your capability to think critically. The more you use such ‘crutches’, the more prominent your ‘mental muscle’ atrophy effect becomes. Even if you manage to produce several pieces of writing for university courses or even several published works, this portfolio will not contribute to your actual understanding of your research area and your skills.
If you’ve used AI-generated content in your dissertation writing, get expert help rewriting today! Our PhD experts can rewrite, edit and add much more in-depth analysis to help you succeed in your PhD studies. Order PhD thesis writing help now.
