Last year, a fellow student submitted their academic paper with a reference that simply said “ChatGPT, 2023.” No link, no version, nothing else. Their professor circled it and wrote “?” in the margin. Fair enough, honestly. Nobody had told them what to do either when something like this happens.

AI went from being a ‘threat’ to ‘unethical’ to finally being somewhat acceptable depending on its use in the world of academia. However, no one knows how to incorporate it into their work. That’s the problem right now. Universities say “declare your AI use” and then leave you to figure out the formatting alone. So in this blog, you will be finding out how the big three referencing styles exactly handle this situation.

Is AI Even Allowed to Be Used in Referencing Essays

Formatting comes second; the first concern should always be whether your institution allows the use of AI for referencing. Some institutions are fine with AI for brainstorming and nothing else. Some want a short statement at the end of your submission. Some ban it for anything assessed. It is important that you look at the guidelines and consult with a teacher first because a neat citation won’t be of any good to you if the tool wasn’t permitted in the first place.

Can AI be Trusted?

Also, a chatbot isn’t a credible and trusted source like a journal article. You can’t verify what it told you, and it makes things up with total confidence. We have seen it invent authors, journals, even page numbers. So if a point matters in your argument, find the real paper or dataset and cite that instead.

This is why many students searching “do my assignment” and pasting whatever the first result comes up in their submission backfires so often. The answer can be wrong; it can get flagged as AI or worse: plagiarism. And when a professor asks you to explain it in a viva, you will have nothing. These concerns have always existed alongside the miracle of AI. Where it appears to have it all sorted for you, it can even backfire at times.

Ways to Reference AI Tools in Different Citation Styles

With changing times, people have learned to accept and integrate AI in their daily lives. Teachers and institutions have become somewhat accepting of the use of AI as long as students are transparent about its use. So here are some ways to reference AI using three major referencing styles:

APA 7th edition

APA’s actually one of the easier ones here. Whoever built the tool is the author; that’s the rule. In-text, that’s just (OpenAI, 2026), reference list looks like this:

OpenAI. (2026). ChatGPT (model name) [Large language model]. https://chatgpt.com

Model name goes in if you’ve got it. APA also wants your prompt in there somewhere too, what you actually asked, comparison, list, whatever it was.

The reader can’t open your chat obviously, so some professors want the full exchange, prompt and response both, stuck in an appendix. And if you’re quoting the output, use actual quotation marks, same as you’d do with any other source.

Harvard

Here’s the annoying bit. Harvard isn’t really one thing. Most UK schools run Cite Them Right, or their own version of it tweaked slightly, so what passes at one university gets marked wrong at the next. A fairly standard version looks like:

OpenAI (2026) ChatGPT [Large language model]. Available at: https://chatgpt.com (Accessed: 2 October 2026).

Some guides want the prompt itself used as the title instead, so you’d write something like ChatGPT (2026) Response to ‘Explain price elasticity of demand’, 2 October. In-text citations are still (OpenAI, 2026) either way. Watch that access date; these tools change so often it genuinely matters. And just check your own library guide before copying any of this blind. Saves marks.

OSCOLA

Law students, this bit’s for you. OSCOLA runs on footnotes, and the fourth edition’s got no official rule for AI at all right now. Plenty of law schools just tell you outright: don’t cite AI for actual legal points; law comes from cases and statutes, not a chatbot’s best guess.

If you still need to reference it somewhere, most people treat it like a web source; the prompt described right there in the footnote:

OpenAI, ‘ChatGPT’ (response to prompt ‘Summarize the neighbor principle’, 2 October 2026) https://chatgpt.com accessed 2 October 2026.

Ask your faculty before you submit. Some prefer a declaration over a footnote, and they won’t mind you asking. And if you still do not know how to reference these tools, for example, for your economics course, getting proper support from Economic assignment help UK experts can help you understand the material for real. Then the work you hand in is yours, and you can defend it when someone asks.

Common Mistakes Observed

Scholars still make several common mistakes, since using AI for referencing is still unacceptable for many. Here are a few mistakes that can be common among students today:

• Citing AI as the source instead of the place the fact came from.
• Copying a reference list the chatbot generated without checking whether the papers are real. Some aren’t.
• Leaving out the prompt, the version, or the access date.
• Switching styles halfway through, usually because someone copied a citation from a different guide. Examiners notice that fast.

A life-changing hack is to keep jotting down the tool name, the query you gave to it, and the date you gave it on. This will save you tons of time and help avoid the panic on the night of submissions

Conclusion

The rules are still changing. Whatever views the academic institutions have on AI today are most definitely going to be different from tomorrow. A helpful tip is to not follow any guides available, including this one, as the final word. Look up your college’s guidelines or what your professor instructed. Just be honest about what you used, check every claim, and build your argument on sources you can actually trust.