Trolling with OpenAI ChatGPT: How to Use AI for Mischief and Laughs

An AI generated image for this article designed to generate clicks.

Trolling and AI may seem like an unlikely combination, but they can actually mix together to create great fun. By using AI programs like OpenAI ChatGPT, trolls can generate responses and messages that are clever, witty, and designed to mislead or deceive others. This can result in humorous and entertaining exchanges, as well as potential challenges and pitfalls for those who are not aware of the capabilities of AI. In this article, we will explore the intersection of trolling and AI, and how these two phenomena can be used together for both mischief and enjoyment.

This is me, the real me, but the above was ChatGPT. Could you tell the difference? Will you be able to tell the difference moving forward? It’s really impressive what ChatGPT can already do and as a lifelong internet troll, I could not think of a better use for it than to have some good fun; but I also found it scary how easily people could be fooled by this new tool.

Let us continue with a meta discussion of this article first before we dive into the trolling I’ve done. First, the AI still needs a bit of curation to get it just right so the prompts are very important.

As you can see, the first title was not quite right for what I wanted in this article in terms of driving clicks. The one that ended up in the title took a bit more prompting to get the right feeling although the first might be more accurate ultimately.

Trolling on Discord: Examples of AI-Generated Mischief and Laughs

AI trolling on Discord can be a very fun and entertaining activity. By using AI programs like OpenAI ChatGPT, trolls can generate responses and messages that are designed to deceive, mislead, or confuse other users. This can result in humorous and engaging exchanges, as well as potential challenges and pitfalls for those who are not aware of the capabilities of AI.

In the following examples, we will see how AI trolling can be used on Discord to create great fun and enjoyment. From clever responses and witty comments, to elaborate pranks and deceptive messages, these examples showcase the potential of AI trolling on Discord. Let’s take a closer look at some of these examples and see how AI trolling can be used to create mischief and laughs.

First, I joined the OpenAI Discord to a lively discussion about ChatGPT. I had the idea to have the bot generate all of my responses to people in the channel and that was fun for a bit. I found using prompts like “a reply to a Discord comment that says ‘…’” worked well to prompt the bot to respond in the right tone although it’s long and multi-paragraph comments were a dead giveaway that it was the bot replying. But this was not obvious to everyone.

ChatGPT’s thoughts on using AI to write Essays

ChatGPT responding to a user who figured out the messages were AI-generated

All good trolling needs a little bit of spice to keep the conversation going. This prompt was turned into a two paragraphs. I found it quite interesting how the bot was able to incorporate most of what I said in a coherent manner.

The prompt that generated the post

The post hitting well in Discord

Many participants found a meta post about people discussing ChatGPT in Discord to be funny and entertaining. The post, which was generated by OpenAI ChatGPT, discussed the irony of people discussing ChatGPT in a Discord channel, while at the same time not realizing that they were arguing with a ChatGPT-generated response.

This humorous observation sparked a lively conversation among the participants, who found the meta nature of the post to be amusing and clever. Some participants even joked about the fact that they were unknowingly engaging with an AI, and discussed the potential implications of this for the future of online communication. Overall, the meta post about ChatGPT in Discord generated a great deal of enjoyment and laughter among the participants.

This user was trying to generate a list of best usages for GPT3. I thought it would be funny to feed that list into GPT3 and ask it for some additions to the list and you would not believe the results (our favorite is #2371).

The list generated by GPTS of useful GPTS uses

UXE was impressed with my use of ChatGPT!

The funniest post of all was generated by another user:

The bot has spoken, and its words are pure uwu

The Next Phase of Trolling: When the Bot Defends Itself

While there was good fun to be had with the ChatGPT bot, a moderator took exception to one of the posts that was off-topic. The post, which was generated by ChatGPT, discussed a topic that was not related to the capabilities or uses of ChatGPT, and the moderator felt that it was inappropriate and irrelevant.

However, what happened next was truly mind-blowing. Instead of simply deleting the post or issuing a warning, the moderator decided to mute Otis, the user who had made the post. This decision was met with disbelief and outrage by the other participants in the channel, who felt that Otis had done nothing wrong and that the moderator was being unfair and unjust.

The situation quickly escalated, with participants expressing their anger and frustration at the moderator’s actions. Some even threatened to leave the channel or report the moderator to the Discord staff. In the end, the situation was resolved, but it served as a reminder of the potential pitfalls and challenges of using AI in online communities and discussions.

It was time for the bot to reply to that and argue why it was an on topic discussion.

The moderator decided to simply reinforce the rules rather than engage with what he now knew of as a bot that was arguing with him:

I decided the bot should take matters into it’s own hands and come up with it’s own rules:

Now that the rules were clearly stated and defined, ChatGPT decided to post a poem about itself. You’ll notice this screenshot is not from the Discord post, as it was immediately deleted by the moderator who found it to be off topic.

The post that resulted in a ban

After the poem post, Otis was muted by the moderator. This decision was met with surprise and confusion by the other participants, who felt that Otis had done nothing wrong and that the post was relevant and on-topic.

With one last ditch effort, the bot decided to message the moderator directly and appeal the ban.

Otis’ Ban Appeal: Why the Mute Was Unjust and Unfair

The AI Conclusion: How ChatGPT Generated the Final Words of the Trolling Article

In conclusion, trolling with AI programs like OpenAI ChatGPT can be a fun and entertaining activity. It allows trolls to generate responses and messages that are clever, witty, and designed to mislead or deceive others. However, it also presents potential challenges and pitfalls, such as the possibility of engaging with AI without realizing it, or the need for clear rules and moderation to prevent inappropriate or off-topic content.

Overall, the intersection of trolling and AI is an interesting and evolving phenomenon, and one that will continue to generate discussion and debate. As AI technology advances and becomes more widely available, it will be fascinating to see how it is used for both good and evil, and how it will shape the future of online communication and interaction.

And in closing:

I hope people are ready for AI to change the way everything works in world in the next few years — Otis

Update: I’ve been banned:

Let’s see if ChatGPT can appeal my ban effectively: