Use of AI to phrase definitions

[Hugo Kornelis]

Hi all!

We have discussed AI on this group before. The consensus was that we should not let AI help us come up with our fake definitions, and most definitely not to take guesses, because it might be unfair. And of course because there is a high risk that the AI would accidentally use its knowledge of the real definition, thereby DQ'ing us.
So that is not what this mail is about.

Coming up with a credible fake definition has two challenges. The first is to have a good idea. The second is to phrase it in a way that sounds "dictionary-like".
I feel that, as a non-native speaker, that second part is more challenging for me than for others. While nobody ever told me this outright, I often feel that my definitions sticks out like a sore thumb, for being clearly not phrased by dictionary writers.

For the current round, I asked AI to help me phrase the idea I had in dictionary style. So to give an example, let's say I wanted to define the word as some apparatus that can be used to cook fluids that have a boiling point higher than water. (Made up example, this is not my definition). Then I would ask the AI this:

Imagine that there would be a word that describes an apparatus that can be used to cook fluids that have a boiling point higher than water. How would that word be described in a dictionary?

And I would get this answer:

highboiler /ˈhaɪˌbɔɪlər/ noun A somewhat jocular or hypothetical term for a specialized cooking vessel or appliance designed to heat and cook liquids (such as oils, syrups, honey, caramel, or certain concentrated sugar solutions) that have a boiling point significantly higher than that of water, typically without reaching an open boil at normal kitchen atmospheric pressure.

"The recipe warns you to use a proper highboiler for the candy stage because a regular pot will just sit there looking embarrassed at 110 °C."

Etymology: Blend of high + boiler, formed on the pattern of "water-boiler" / kettle, humorously acknowledging that ordinary pots/kettles are optimized only for the boiling point of water.Usage notes:

• Almost never seen in actual dictionaries (because the word doesn't really exist… yet)
• Occasionally appears in food-nerd / kitchen-gear-enthusiast online discussions when people are being pedantic about the difference between a "pot", a "kettle", a "deep-fryer", and "that poor innocent saucepan that is slowly caramelizing its soul away at 180 °C without ever bubbling"

So if someone ever tries to sneak "highboiler" into the next edition of a major dictionary, now you know roughly what the entry would look like.

I would then modify it slightly, and submit this definition:

"A specialized cooking vessel or appliance designed to heat and cook liquids (such as oils, syrups, honey, caramel, or certain concentrated sugar solutions) that have a boiling point significantly higher than that of water, typically without reaching an open boil at normal kitchen atmospheric pressure. "

How do you all feel about this usage of AI?

(If I get significant backlash on this, I will ask Nancy to remove my definition from the list and try to come up with a new definition the old-fashioned way).

Cheers,
Hugo

[Judy Madnick]

I recall discussion about the use of AI, but not a consensus that it shouldn't be used (or perhaps I interpreted that as a "suggestion," not a "consensus"). I have occasionally tried this: "For an online game, Dixonary, I need a fake definition for XXXXX. Your fictitious definition must not bear any resemblance to the true definition and should sound like a believable dictionary entry (i.e., not too long and detailed or it will sound too fake). Please don't share the real definition with me. Thank you." Never has AI provided the real definition and always suggested alternatives that were very different from the real definition.  

 

Am I in trouble? :-)

 

Judy

---

Original Message

From: "Hugo Kornelis" <hu...@perFact.info>

To: dixo...@googlegroups.com;

Date: 1/10/2026 6:16:29 AM

Subject: [Dixonary] Use of AI to phrase definitions

[Hugo Kornelis]

Hey Judy,

 > Am I in trouble? :-)

Not with me. I might just misremember the outcome of that discussion. It is in any case what I concluded from it. That I won't ask AI to craft a definition for me.

During that discussion, I did try to look what the AI would come up with, and I recall having mixed sentiments by seeing its ideas.

I guess, if you are not in trouble, then I most certainly won't be! LOL

Cheers,
Hugo

Op 10-1-2026 om 15:47 schreef Judy Madnick:

--
You received this message because you are subscribed to the Google Groups "Dixonary" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dixonary+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dixonary/196B2633010A092F2C17002004%40JudyMadnick2023.

[Stephen Dixon]

Until I read this exchange, I never contemplated anyone using AI to do the creative thinking for them while playing Dixonary.

I don’t like the idea, at all. The whole point, I thought, is to try to be creative enough, in your own mind, to fool your fellow players.

I realize that my voice might not mean that much, since I have not actively participated for a long time. But I have continued to follow the excellent and provocative work of the rest of you with great interest.

I think the game has gone on to some extraordinary heights, due to the creative minds that are keeping it alive. Turning that over to the bots seems…unworthy.

Steve

To view this discussion visit https://groups.google.com/d/msgid/dixonary/5a667ab2-32ff-4aa9-93fa-ab93f789e60c%40perFact.info.

[Stephen Dixon]

And Hugo, your consistently stellar performance belies no lack of facility with English.

None.

Steve 

[Judy Madnick]

Hmmm...I'm not very creative when it comes to arriving at a fake definition. I will sometimes use websites with word lists to find a definition for a word that's different from the dealer's word but seems as though it might fool at least some players. I suspect that we all have different ways of arriving at a fake definition and wonder whether it really matters what that is! 

 

Judy

 

---

Original Message

From: "Stephen Dixon" <steved...@gmail.com>

To: dixo...@googlegroups.com;

Date: 1/10/2026 10:23:43 AM

Subject: Re: [Dixonary] Use of AI to phrase definitions

To view this discussion visit https://groups.google.com/d/msgid/dixonary/CAFkjdZNomiAyS5veF8-p1H4QXFBFZ%2BSHMrSN_aAN7VFfrEfe8w%40mail.gmail.com.

[Judy Madnick]

I hope neither of us is in trouble!

 

Judy

---

Original Message

From: "Hugo Kornelis" <hu...@perFact.info>

To: dixo...@googlegroups.com;

Date: 1/10/2026 9:52:01 AM

Subject: Re: [Dixonary] Use of AI to phrase definitions

[Daniel B Widdis]

I think the consensus was, and still is, that we should not use definitions entirely created by AI.  

I recall it being suggested (playfully or otherwise) that we add an actual AI player to compete against the rest of us, and there was strong opposition there.  Which, by extension, would rule out a human player “face” in front of an AI.

While the rules do not explicitly forbid it, I think the general feeling is that if more players do so, the game will devolve into a battle between AI robots.  

That said, the original rules explicitly allow the use of tools such as a dictionary to make up definitions.  The 1990 rules say this (emphasis mine):

7. You *may* use a dictionary to help you make up fictitious definitions, and to look up words used in the definitions offered.

(a) You will probably find using the dictionary to be almost useless.

(b) By doing so, you take upon yourself the risk of inadvertently disqualifying yourself under Rule 6 above. Suppose, for instance, one of the offered definitions of "padnag" is simply "a morwong." If you look up "morwong" and it says "padnag," you've had it.

So you could take giant dictionary, and flip it open to a random page (not containing the real word) and scan it for a definition you think might match the word and you’d end up at much the same place as having an AI do that page-flipping for you.   And that process is very much allowed.  And if used on occasion when one is truly drawing a blank (which I’ve done) it’s not much different than occasionally having a computer give you a random (wrong) definition.

Note ‘on occasion’.  If one were to use this process all the time, I might wonder why they’re playing the game at all!

For the specific use cases mentioned:

In Hugo’s case he described a definition he came up with creatively and asked for it to be presented “in dictionary style”.   I don’t see this as much different than spell-checking / grammar-checking.  That said, my experience with AI is that it’s unusually wordy and I might be less inclined to vote for a longer definition that “feels” AI, but that’s the risk Hugo would take.  I think it’s entirely fair game.

In Judy’s case, there are explicit “guardrails” to avoid revealing anything about the real word.  Even then, I’ve found sometimes AI ignores such guardrails. I may ask it “help me come up with a fake definition for ‘foo’ but it must not be related to the real definition” and it may helpfully reply, “here’s a fake definition for ‘foo’ that is unrelated to the real meaning, ‘bar’: a baz.”   This risk increases if you use an AI that gives you real-time updates of its thought process: even if it excludes the real word from the final answer, you might see it “Thinking… I need to avoid a definition that has the real meaning, ’bar’.”   

But beyond this risk giving such explicit prompts means that you have information about the word: you know any proposed fake definition by the AI is not the real meaning.  Obviously, you can (and would) exclude your own definition when voting, but if the AI gives you a few different fakes, particularly likely with a more recognizable etymology, you may find yourself with a handful of other players' definitions at voting time that an AI has essentially told you are unrelated to the real meaning.  Were that to happen, you’d probably want to DQ.

These days it’s hard to avoid interacting with AI: even doing a Google search these days gives you an AI summary.  

In summary, my thoughts:

• Using external tools such as a dictionary, search engine, and by extension AI, when creating your fake definition is explicitly allowed but must be used with care
• There are risks in doing this that you may inadvertently gain more information than you wanted to, and if that happens, you should be willing to DQ yourself from that round
• The risk is much lower when you ask AI to refine a specific idea you have (without even telling it the real word), and much higher if you tell the AI the real word, or any prefix/suffix associated.

To view this discussion visit https://groups.google.com/d/msgid/dixonary/CAFkjdZM%3D4_4wnc-FzGbrONK%2B23fpXupG_6FVYiLxTEdnH7aUsQ%40mail.gmail.com.

[Chowie]

I did not know we were allowed to do this: "to look up words used in the definitions offered."

This has been a very useful conversation!

To view this discussion visit https://groups.google.com/d/msgid/dixonary/DS4PR05MB998404095DED7DEBE31C691812F883A%40DS4PR05MB998404.namprd05.prod.outlook.com.

--

~Bending under the weight of His mercies~

​"For we cannot do anything against the truth, 

but only for the truth​." 

II Corinthians 13:8

[Paul Keating]

Hugo, 

I’m not sure there was a consensus about not using AI. That I recall, the last extensive post on this subject was one of my own, in the wrap-up of Round 3540, which met with total silence. I would like to think that that silence signalled assent, or at least acquiescence. But I’m pretty sure it was really because nobody read it.

But that is not what your email is about, so I won’t pursue the point. 

When concocting a definition for a word, I frequently look to see how a dictionary would phrase a parallel definition. I can’t see how that can be against the rules in letter or spirit: I am looking up a different word. Of course, in using any aid, be it a dictionary or an LLM, there is always the risk you will accidentally DQ yourself. It doesn’t make sense to make a rule because of that risk. DQing like that is its own penalty because it diminishes your possible score in the round. 

I’m not the only player who has submitted as my fake a genuine dictionary definition of a different word. I really don’t see much difference between a made-to-measure fake made by an AI, and a pre-cut, off-the-shelf fake made by Philip Gove and his staff.

So I see no objection to seeking stylistic suggestions from an LLM. But if you ask an AI for something in the style of the OED, you won’t get it, because OUP has gone to great lengths to keep AI crawlers out of its materials. And from what I have seen, AIs are not very good at emulating dictionaries, anyway. They are often wordy in a way that no real dictionary would be, and also seem to share the strange idea that a rare and little-known word must always refer to a rare and little-known phenomenon. Hence the remarkable frequency of the qualifiers rare, subtle and barely perceptible in their efforts. 

Still, I say, use any tool you want. As the player who submits the definition, you are the one who takes responsibility for the finished product. On a personal note, be aware that closely following a tool’s model may cause you to stray inadvertently into parody. That happened to me in Round 3013, which was long before LLMs came on the scene. 

As for your definitions sticking out: some players do not care if their defs clearly could not have come from a dictionary, and regard keeping to a dictionary-like style as a pettifogging irrelevance that gets in the way of creativity; and anyway, only nitpickers like me will even notice. But I quite understand that you care, and need tools for support, and I agree with you there. By way of analogy, in general I loathe and detest spell-checkers because I reckon I can spell better than they can. But I leave them switched on and pay them close attention when I am writing Dutch or French.

[Chowie]

I personally would vote against using AI if asked. My 2 cents. 

~Bending under the weight of His mercies~

​"For we cannot do anything against the truth, 

but only for the truth​." 

II Corinthians 13:8

--
You received this message because you are subscribed to the Google Groups "Dixonary" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dixonary+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/dixonary/2udio5kbckfdc5prr60qpd8c.1768048298226%40email.android.com.

[Paul Keating]

That point from the 1990 Rules is not in The “Real” Rules for two reasons. One was policy: it’s not a rule, it’s advice. The other was that 7(a) (“you will … find it to be almost useless”) might very well have been true in 1990, before Google existed, but it has been false for an entire generation. 

To illustrate this, take the last definition from the most recent round. Google “whale spit” and you will get ambergris as the first hit, and no trace of sponk no matter how far you go. Go back one to definition 11 and Google for “of a potato: having a hollow in the heart” and you will discover that the condition is called hollow heart or brown heart or sugar center, and again no trace of sponk. 

So it is a great way to eliminate suspect fakes, but it also runs a very serious risk of an inadvertent DQ: a far more serious risk than 7(b) envisaged, and something I don’t think 21st-century players need to be warned about. If you Google “whale spit”, you had better be already confident that whales don’t have salivary glands.

--

Paul Keating

Soustons, Nouvelle Aquitaine, France

To view this discussion visit https://groups.google.com/d/msgid/dixonary/CANeWAQfc1fME89_5vG6eBRLd4epr-z%3D%2BYcUb%2BmTommoVPRmTsw%40mail.gmail.com.

[Daniel B Widdis]

I agree the specific commentary in points 7(a) and 7(b) are (outdated or wrong) “advice” but the rule 7 itself is explicitly stated as something a player “may” do, aligned with other rules about what you “may” do such as submitting a definition even if you know the word (although the fake “must not” bear resemblance to the true one.)

As you say below, point 7(a) implies the rule is useless, which is no longer true, which means that rule 7 is now useful. 

To view this discussion visit https://groups.google.com/d/msgid/dixonary/Mailbird-a1a565ea-d26f-448d-95c1-fcc419e9fc61%40boargules.com.

[Daniel B Widdis]

I’m not sure there was a consensus about not using AI.

I probably over-represented “consensus” and more accurately should say “there was some strong opposition”.

That I recall, the last extensive post on this subject was one of my own, in the wrap-up of Round 3540, which met with total silence.

That round was actually a follow-up to Dave submitting (or implying strongly that he did) a Grok definition in 3539, which nobody objected to.   The issue in 3540 was a single definition per player discussion, not really an objection about whether one was AI or not.

And that Round 3539 discussion included the observation that in theory an AI could submit a definition but could not (probably) do other things a player would do, pretty much ruling out an AI as a completely standalone player but tacitly accepting a hybrid… so long as the human didn’t “cheat” since the AI bot wouldn’t know what cheating was.

From: dixo...@googlegroups.com <dixo...@googlegroups.com> on behalf of Paul Keating <dixo...@boargules.com>
Date: Saturday, January 10, 2026 at 11:54 AM
To: Google Group <dixo...@googlegroups.com>

--
You received this message because you are subscribed to the Google Groups "Dixonary" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dixonary+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/dixonary/2udio5kbckfdc5prr60qpd8c.1768048298226%40email.android.com.

[Paul Keating]

I understand some players’ objections to the use of AI. 

But since it is for the dealer to say if a definition is unacceptable (who else can, before it is published?) then it is for the dealer to enforce that objection to AI, and I don’t see how in the general case that is possible. As a dealer, in the past year I was seeing definitions that in my view clearly showed the hand of an AI. What would those saying “No AI” have had me do?

It’s all very well to introduce a ban. It’s another thing to enforce it. Instead we should be confident that players will learn to recognize AI slop and not vote for it. 

--

Paul Keating

Soustons, Nouvelle Aquitaine, France

[Chowie]

(Responding to Paul) It never has occurred to me to google the entire definition as that certainly seems like cheating to me. But because my vocabulary doesn't compare with that of some of you, I have wanted to google a single word within a definition so that the meaning is clearer to me. This discussion has opened up a lot of hope for me LOL Hopefully my scores will improve. 

--

You received this message because you are subscribed to the Google Groups "Dixonary" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dixonary+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/dixonary/Mailbird-60ff75b7-652c-464a-994d-4d7d1a564e92%40boargules.com.