Elon Musk's Grok is making it easier to dunk on someone

• People are replying to posts on X — asking Grok: "Is this true?"
• It's a new form of dunking on someone on X, and it's making the old Twitter an even weirder place.
• There's a limit to how useful a chatbot like Grok can be on social platforms — and we're seeing it.

Lately, I've noticed something surprising on X: People are actually using Grok. Like, a lot.

Before I describe why this is kind of weird — and making X kind of weird — I should give a disclaimer that I generally find X to be worse than it used to be in both specific and vague ways.

I don't think Grok is to blame; it's more the boring reasons you've probably heard before: a lot of the lively and funny users I used to enjoy have quit the platform, the remaining users seem to be more combative, the "For you" feed is dominated by clickbaity viral video chum, there's a lot of spam and porn.

The thing that I've noticed outside all that stuff is this new phenomenon of asking Grok questions in a reply to a post. Typically, a person is asking Grok — Elon Musk's ChatGPT competitor —to explain the context of a tweet or to verify if something in the tweet is true.

Here's a made-up example of the kind of stuff I see:

@PersonA: The moon died because it was vaccinated

@PersonB: @grok, is this true?

@grok: The moon is not alive and, therefore, cannot be vaccinated.

When I describe it like that, it sounds very normal and useful, right? Well, in practice — as with a lot of things connected to LLMs — it sort of makes things … just weird.

How Grok responds

I've spent some time browsing the "Replies" tab of @Grok, which is a fascinating firehose of all the replies to various things people are asking it. (Note: I'm talking about publicly asking @Grok as a reply to a post on X — not the private conversations you can have with Grok.)

What I tend to see a lot: people responding to a political post with some sort of fact-check. The topics tend to be the kinds of things that some people might want to stir up fights about: Trump's policies, the history of Jews, immigration, Elon Musk, and a lot of questions related to Hitler. (There's also a lot of stuff that isn't in English, so I can't say what that's about.)

One recent "fact-check" came in response to a post from Musk himself:

(For this meme, Grok said no, it wasn't factual. I asked the person who runs the @Dispropganda account if they knew in advance they were going to get that answer, and they said, "I didn't know for sure, but I suspected.")

In theory, this is good, right? Finally, people are able to get fact-checks on posts and access to unbiased opinions!

But what Grok seems to be used for more often is a rhetorical cudgel, a debate tactic. It's not being used for earnest requests for information but as a new tool in the dunk utility belt.

I've seen people use it on posts they know aren't true — and are purposely getting Grok to point that out.

Asking Grok to weigh in on someone's tweet has its own passive-aggressive vibe. Instead of engaging the original poster, you're appealing to this automated thing.

The fact that people seem to be using it during debates on heated topics where people generally have entrenched ideas suggests that people aren't using it as a helpful resource during lively discussions. They're using it to fight and argue, to insult the other person. They're using Grok as an AI dunk assistant.

And instead of elevating the level of discourse on X with more truth and utility, it's just dragging it down with more nasty arguing and petty trolling. (X didn't immediately reply to a request for comment on this story.)

AI has its limits on social media

I'm not a total AI skeptic — I believe there can be some amazing things happening quite soon. But as far as getting an AI bot to engage with you on social media? I really think we're already seeing the limits here. It might be mildly useful in some limited instances, but it's not making the overall X experience better and certainly, in some cases, is making it slightly worse.

I don't mean to be sentimental, but part of the beauty of a text-based social platform is that you CAN just reply to people and engage with them, and, yes, even argue with them! Or if you see something that needs context, you can ask for it or just keep scrolling.

I saw someone ask Grok to explain the context of some deep stan lore tweet about Selena Gomez and Justin Bieber, and I kind of felt sad in a way. Back in MY day, you needed to actually spend tons of time watching the strange motions of the Seleners to be able to understand their garbled system of beliefs — and there was some beauty to that.

Now, you can have Grok — or another chatbot — just spit it out. I'm not sure that's a totally good thing.