Generative AI tools can significantly assist and enhance efficiency in learning and research. However, there are important things to keep in mind when using the results generated by these tools!
While many AI tools draw from vast
amounts of reliable data, their "black box" nature can make it difficult to verify
sources. AI-generated content may be biased or inaccurate due to training data
disparities (e.g., more English data than Chinese, or more data from other regions
compared to Taiwan). If you cannot verify the sources of generated content or the
training data, you cannot guarantee its accuracy or appropriateness. Always maintain a
healthy skepticism and independently verify the information.
[What you
can do]
Consider using multiple, varied sources to cross-verify the data
obtained from generative AI tools.
Generative AI tools are trained on
vast amounts of data, but the legality of how this data was acquired is often debated.
While these tools boost efficiency, their outputs may include unauthorized citations or
copyrighted material. There are currently many lawsuits accusing AI companies of
copyright infringement.
As copyright ownership of AI-generated content remains
unclear and fair use principles are unstandardized, academia is actively developing
guidelines. Users should carefully verify results and appropriately disclose their use
of AI tools. Adhering to academic integrity is crucial for responsible research and
mitigating legal risks.
[What you can do]
① Carefully verify
generated results and disclose the extent to which AI tools were used.
② If
submitting an article for publication, check the publisher's policies and restrictions
on AI usage beforehand.
While AI tools can help explore
topics, they are not academic databases or standard search engines. Even if trained
on academic literature, generative AI is designed to "generate" new text, unlike
databases that point you directly to primary literature and sources. Keep this key
difference in mind.
[What you can do]
① Rely on actual
academic databases or library discovery platforms when searching for peer-reviewed
literature.
② Consider using AI tools developed by academic publishers (like
Scopus AI), which are trained on their own verified databases and are far less
likely to hallucinate or provide false information.
Because different AI tools use
different language models, the recency and accuracy of their information vary. When
searching for current events or recent developments, an AI tool is only as good as the
cut-off date of its training data.
For example: On February 3, 2025, ChatGPT-4's
latest training data was up to June 2024. Therefore, it would be unaware of any events
between July 2024 and February 2025. On the same date, Microsoft Copilot had data
updated up to January 23, 2025, offering a much more recent
perspective.
[What you can do]
If data recency is critical to
your research, use frequently updated academic databases like PubMed, Scopus, or Web of
Science.
Everything you input into an AI tool
might be used as future training data and could inadvertently resurface in responses to
other users. You must handle personal and confidential information carefully to prevent
data leaks or unauthorized access.
[What you can do]
① Be
extremely cautious about what you type or upload. Never input sensitive, confidential,
or proprietary information.
② Familiarize yourself with the privacy policies of the
AI platforms you use to ensure compliance with academic ethics and data protection
regulations.
Reference Guidelines for the Use of Generative AI by the Executive Yuan and its Subordinate Agencies
Ethics of Artificial Intelligence: The Recommendation
Guide on the use of generative artificial intelligence
Generative artificial intelligence (AI) in education
Reference Guidelines for Teaching with Generative AI Tools [1st Edition]
How to cite
ChatGPT
→ Guidelines and Examples https://apastyle.apa.org/blog/how-to-cite-chatgpt
Key
Points
★ AI cannot be listed as an author in APA academic
publications.
★ Authors are fully responsible for the accuracy of any
information in their papers.
How do I cite generative AI
in MLA style?
→ Guidelines and Examples https://style.mla.org/citing-generative-ai/
Key
Points
★ You must cite the AI tool whenever you paraphrase, quote,
or incorporate any AI-generated content (text, images, data, etc.) into your work.
★
It is not recommended to treat AI tools as authors.
How to cite content
generated or developed by AI (e.g., ChatGPT)
→ Guidelines and
Examples
https://www.chicagomanualofstyle.org/qanda/data/faq/topics/Documentation/faq0422.html
Key
Points
★ Do not cite ChatGPT in a bibliography or reference list
unless you provide a publicly available link to the conversation.
AI Website Generator
