https://www.explainxkcd.com/wiki/api.php?action=feedcontributions&feedformat=atom&user=2A09%3ABAC2%3AC729%3ADE1%3A0%3A0%3A162%3A36explain xkcd - User contributions [en]2026-04-17T03:11:32ZUser contributionsMediaWiki 1.30.0https://www.explainxkcd.com/wiki/index.php?title=3126:_Disclaimer&diff=3837993126: Disclaimer2025-08-09T01:33:00Z<p>2A09:BAC2:C729:DE1:0:0:162:36: /* Explanation */</p>
<hr />
<div>{{comic<br />
| number = 3126<br />
| date = August 8, 2025<br />
| title = Disclaimer<br />
| image = disclaimer_2x.png<br />
| imagesize = 346x396px<br />
| noexpand = true<br />
| titletext = You say no human would reply to a forum thread about Tom Bombadil by writing and editing hundreds of words of text, complete with formatting, fancy punctuation, and two separate uses of the word 'delve'. Unfortunately for both of us, you are wrong.<br />
}}<br />
<br />
==Explanation==<br />
{{incomplete|If the academic integrity department asks, this page was NOT created by ChatGPT. Don’t remove this notice too soon.}}<br />
{{w|ChatGPT}} is a large language model, a type of bot commonly marketed as {{w|generative artificial intelligence}}. In 2025, at the time of this comic’s release, it was common for people to use ChatGPT to generate emails and forum posts. Since ChatGPT-generated texts are often very wordy, a side effect of this phenomenon is that people who actually write long messages are sometimes incorrectly accused of using ChatGPT to generate them. Accusers may also be suspicious of messages that use formally-correct punctuation such as {{w|Dash#Em_dash|em dashes}}, which most online posters don’t know how to type and which ChatGPT uses sometimes.<br />
<br />
In the comic, [[Cueball]] (possibly representing [[Randall]]) has written what appears to be a forum post of some kind, from the juxtaposition of user-profile image and information against it. At the end, possibly as an automatic forum signature, he includes a notice disclaiming the idea that the message might have been generated by ChatGPT.<br />
<br />
The title text expands on this joke. Randall claims people say that no human would write a thorough essay with bullet points and the word “delve” (a word not commonly used by human English speakers and overused by ChatGPT) on {{w|Tom Bombadil}}, a minor {{w|Lord of the Rings}} character, and Randall says they are wrong because he would and did. The reason Randall used the word “delve” may be that he was discussing a line from ''The Fellowship of the Ring'' where the character Gandalf describes the Dwarves as having once “delved too greedily and too deep”. It could also be a reference to "Michel Delving," the largest Hobbit town in the Shire, whose name means "Large Excavation." <br />
<br />
This may be related to GPT-5, which was released a day before this comic was created.<br />
<br />
==Transcript==<br />
{{incomplete transcript|Don't remove this notice too soon.}}<br />
:[A Cueball profile picture is shown next to five paragraphs of illegible text, which contains varied punctuation marks and two square-bracketed citations.]<br />
:[The last paragraph in the essay is separated from the other paragraphs by a line with three dashes. Its single sentence is highlighted, and lines connect that illegible sentence to a box with an enlarged, legible version of the sentence.]<br />
<br />
:Enlarged text: Not ChatGPT output—I’m just like this.<br />
<br />
:[Caption below the panel:]<br />
:I’ve had to start adding this disclaimer to my messages.<br />
<br />
<br />
{{comic discussion}}<noinclude><br />
[[Category:Comics featuring Cueball]]<br />
[[Category:Artificial Intelligence]]<br />
[[Category:LOTR]] <!-- title text --></div>2A09:BAC2:C729:DE1:0:0:162:36