The Ethical Dilemma of Utilizing AI in Academic Research
Debating the Ethics of AI in Scientific Research
I recently had a discussion with a PhD candidate about the ethics of using AI in scientific research. The debate revolved around the thin ethical line that exists in utilizing AI models. The question remains, how are standards set for the use of AI in academia, and are they unbiased or universally agreed upon?
AI models have become known for their ability to systematically insert predicted words into specific sequences, creating the illusion of human-generated content. However, these models are not truly autonomous as they are trained on billions of parameters.
A basic prompt will result in a desensitized response, or even a repeated set of words phrased in different sequences. Unless the prompt is crafted properly, the model will not perform well, like a child struggling to assemble words to form sentences.
Navigating the Challenges of Paper Writing
In academia, researchers strive to write impactful publications by performing extensive literature review and conducting experiments. The goal is to advance in their field, provide value, and be recognized as a leading expert. However, the process of writing a paper is not easy. After the research and testing, the real challenge is writing the paper in a way that captures the reader’s attention, presenting ideas in a systematic and engaging manner. This requires a high level of proficiency in the language used, proper citation, and an active writing style.
The Conventional Editing Process
In the editing stage, some researchers often turn to tools like Grammarly to improve their writing and perfect the language. Most researchers conduct internal reviews before submitting the paper for peer-review or other forms of review.
The external review process can take several months and involves continuous editing of the paper based on the reviewer’s feedback, fixing data, and refining the structure. Once the editing is complete, the paper is ready for publication and potential citations.
AI Editing in Academia: Balancing Authenticity and Ethics
The use of AI in editing raises important ethical questions. There are two scenarios to consider: (1) an authentic author using AI to enhance the quality of the paper similar to a human reviewer, and (2) an unethical author who creates a paper using AI without original thought or contribution.
The challenge is to determine where the line should be drawn between these two scenarios and ensure that the use of AI in editing upholds the values and standards of academic integrity.
Scenario 1: Using AI for Editing Assistance
In an article published by Nature, the author used GPT-3 to revise their paper, keeping the data, information, citations, and tone unchanged but requesting restructuring and grammar correction. This raises the question: if using Grammarly to correct grammar is acceptable, why not use AI for similar purposes?
AI has the potential to provide additional information that could affect the tone of the paper and inspire the author in new ways. The possibilities are endless. However, it is important to ensure that the use of AI aligns with the standards and values of academic integrity.
Scenario 2: Ethics of AI-Generated Research Papers
Another perspective is that utilizing AI to generate research papers is unethical because the author is not the sole creator of the content. However, our own work is often influenced by existing data and research. Is it unethical to be inspired by others’ work?
The main difference here is the capacity and speed of computation and retrieval of information. AI can process and search data at a much faster rate than humans. However, using AI to generate research papers requires a certain level of expertise in crafting the prompt, editing the output, and demonstrating the validity of the research through experimentation. At minimum, the use of AI in the creation of a research paper should be disclosed.
Acknowledgments in Academic Writing
One issue raised is the question of acknowledgment in academic writing when researchers seek help from their colleagues. While some authors acknowledge their contributors in their papers, others do not. How should this situation be evaluated? Is it unethical to use someone else’s intellectual property without proper citation?
There is no clear answer as the norm is to reference, but there is no strict proof or obligation to do so.
The Irony of AI Assistance
The debate over the use of AI in research and editing continues to evolve. Ultimately, it comes down to the honesty and integrity of the author in acknowledging their use of AI assistance.
However, the irony lies in the fact that while we embrace AI as a tool to aid us, we still want to limit its use in certain situations. The creation of tools to detect AI-generated content only highlights this, as we are developing ways to restrict AI while it has yet to reach its full potential.
Conclusion
Drawing a line between ethical and appropriate use of AI is a challenging task. AI has the potential to greatly enhance many aspects of our lives and technology. For example, it can be used to improve visual learning, improve the lives of individuals with disabilities, track intellectual property and copyrights, detect malicious diseases, enhance cybersecurity, and more.
However, the use of AI raises important questions. Should we be transparent about our use of AI and acknowledge its role in our work? These questions will continue to be explored as we gain more knowledge about this rapidly evolving field.
Join Us Today
MetaEngineers Hub Inc., is a web 3.0 tech startup that is focused on creating an ecosystem where tech professionals from around the world can collaborate, develop, showcase, and sell their innovations while protecting their intellectual property through a decentralized, AI-blockchain driven ecosystem.
If you like our articles, subscribe to our page or turn on your notifications 🔔.
You can also book a meeting with our founder if you are interested in learning more about our vision and what we are building.
