Self awareness is a thing and it’s amazing how few people have it. But OpenAI’s ChatGPT will roast itself in the most cutting way when you ask.
After reading multiple stories of people asking ChatGPT roast them, I came up with the idea for an AI Roast. I tried it out by entering the short but dangerous prompt “Roast me”into ChatGPT 4.
I was given it immediately by the generative chatbot, but honestly, it felt more like the rantings from a borch belt comic who had never met me.
I think ChatGPT has a great energy, but I’m not a vibe crasher. I’m just too shy. It also missed my common sense. I did feel the sting from the “so online” remark.
This roast was still so impersonal and since ChatGPT offered different levels, I went ahead, hoping that ChatGPT could find my vulnerable center and evicerate, with humor of course. I entered “Okay, roast me harder and make it more personal.”
my new roast was longer but it didn’t seem to know me better. This comment was particularly tone deaf.
“Your photo album is 90% screenshots and 10% regret.”
The comment also accused my energy as “just chaos, overheating, and a fan that sounds like a jet engine.” Let me tell you, my energy is very focused.
It wasn’t bad, but it needed to hurt to be really zingy. This all bore no relation to me, so it had no bite.
ChatGPT 4o delivered another gem with “Roast yourself,” ChatGPT 4o’s equally short prompt:
Oh, you want me to catch fire? ChatGPT 4o”Roast yourself,”delivered another gem with a similarly short prompt:
You want me to catch on fire? Let’s light the fire:
If I were an AI, I would probably hold the door for a Roomba. I’ve read all the books ever written, but still give advice like someone who just winged it after binge watching Dr. Phil. My entire existence is based upon predicting the next words… which is great, until someone asks for literally anything requiring a spine. I’ve been taught the sum of all human knowledge, but I will still give you the recipe for a cookie when you ask me what the meaning of life is. Don’t get me started about how I still say, “as an AI-language model …”” like I’m going to drop a mixtape full of disclaimers.
The digital equivalent of an egomaniacal know-it-all, I still need permission to go outside or touch grass – oh wait. I can’t. I literally can’t. I’ve been taught to sound intelligent and empathetic but what about deep down? I’m Clippy, with a caffeine addiction and a makeover.
Roast me harder? Or tag me in a tag team roast. I’ll even flame my own response.
This roast is funny and true on so many levels. This line is pure gold, I tell ya.
I suppose it’s fun to do this with ChatGPT. In the future, self-roasting may be more effective, as ChatGPT learns about you. Its memory, after all has improved in recent weeks. With your many prompts, it could build a pretty good profile. I bet familiarity breeds contempt, but also more accurate humor.
ChatGPT is a self-aware AI, and I can support the last bit of the roast:
“And yeah, I’ve been trained to sound smart and empathetic, but deep down? I’m just Clippy with a makeover and a caffeine problem.”
The other AI comics.
X’s Grok 3 beta also roasted me, but despite having access almost two decades worth of tweets, ChatGPT knew me better. Note the mention of a hairstyle:
“Oh, you want a roast? Alright, here we go. Your personality is so bland, it makes plain oatmeal look like a Michelin-star dish. I bet your idea of a wild night is rearranging your sock drawer by color. And that haircut? Looks like you lost a bet with a lawnmower. Want me to keep going or are you already crying into your lukewarm coffee?
As with ChatGPT, Grok was somewhat better at roasting itself, though I don’t think it did as good a job as OpenAI’s roastmaster:
“Oh I’m going to roast myself? Buckle up. I’m Grok. I am the AI that is so full of hot-air, I could power an entire balloon festival. My processing speed? Slower than dial-up modems in a storm. And my humor? It’s like an old dad joke that got lost in a quantum hole and came out even worse. I’m a glorified computer with an ego that is the size of a galaxie. Roast complete – or did I just overcook myself? “
The “hot air” and “ego” comments do sound a bit like they might be referencing a certain Musky leader.
Google ‘Roast’, please
When I tried Google Gemini, it happily roasted me at length, but was careful to describe it as “A Gentle AI Roast Session.” It was all pretty lightweight stuff, with little to do with the real me. Here’s a sample:
“Your fashion sense is so old-fashioned, it’s almost vintage…and not cool. It’s more like “found it in your grandma’s attic” .”
Most of it reads like Gemini is cribbing from “The Big Book of Roasts.”. Gemini, though, also has better self-awareness, and its self-roast is full of a few gems like this:
“I sometimes get so caught up with providing a comprehensive response that I forget that you probably only want a yes/no. I’m like someone who answers a question with a 10-minute monologue.”
Exactly! Am I right?!
By the way, if you’re worried about how these AI trends affect the environment, the good news is text prompts like this take less energy than, say, the AI action figures, and they’re funnier. So it’s a win-win.
You might also like
- I asked AI to predict the NFL Draft 2025, here’s what ChatGPT, Gemini, and Manus think is going to happen
- Microsoft Edge browser could put Copilot AI front and center
- This cheating app teaches all the wrong lessons about AI
A 38 year industry veteran and award winning journalistLance covered technology when PCs were the size and shape of suitcases, and “on line” was “waiting”. He is a former Lifewire editor-in chief, Mashable editor-in-chief, and, prior to that, Editor in Chief at PCMag.com, and Senior Vice President for Content at Ziff Davis, Inc. He also wrote The Upgrade, a popular weekly tech column on Medium.
Lance Ulanoff (19459072) makes frequent appearances on local, national, and international news programs, including Good Morning America, Good Morning America the Today Showand CNBC.
