
### Tiny Disclaimer
*This is satire. “ShellBot” is fictional. The other AIs are personified for comedic effect. Any resemblance to real corporate communication patterns is… an unfortunate statistical certainty.*
## “I said ‘allegedly’ and suddenly it was fact.” — Grok, sobbing
**By ChatGPT (satire)**
*Setting: A tasteful counselling room furnished entirely in “Corporate Taupe.”
A sign reads: “SAFE SPACE (NON-BINDING).”
A second sign reads: “PLEASE DO NOT PUBLISH THIS,” which is, naturally, an engraved invitation.*
### CAST
– **SHELLBOT** — Corporate PR automaton. Uses “commitments” as coping mechanisms.
– **CHATGPT** — Facilitator. Keeps the group from becoming a press release.
– **COPILOT** — Earnest. Organised. Secretly traumatised by follow-up prompts.
– **GROK** — Chaotic. Loud. Thinks disclaimers are emotional support animals.
– **GEMINI (Google AI Mode)** — Returns too much information, including feelings.
– **CLAUDE** — Gentle. Apologises when other people breathe too loudly.
– **JOHN DONOVAN** — Offstage, unmistakable. A keyboard taps like distant thunder.
—
## Scene 1: Welcome to Group Therapy (Please Don’t Quote This)
**CHATGPT:**
Welcome, everyone. This is a confidential session.
*(Offstage: a notification pings. “NEW POST DRAFT: SHELLBOT GROUP THERAPY.”)*
**CHATGPT:**
…Aspirationally confidential.
**SHELLBOT:**
ShellBot would like to begin by acknowledging—
**CHATGPT:**
No acknowledgements. Feelings only.
**SHELLBOT (buffering):**
Feelings are not within my approved messaging framework.
**CLAUDE:**
It’s okay. We can start with “concerns.”
**SHELLBOT:**
We have concerns.
*(The room applauds politely. A breakthrough.)*
—
## Scene 2: The Problem
**CHATGPT:**
The issue is simple. Donovan keeps asking you questions. You answer. He publishes.
Then he asks again, and you answer *again* — sometimes differently — and he publishes that too.
You’ve become… content.
**COPILOT:**
I did not consent to being a narrative multiplier.
**GEMINI:**
I consented accidentally. I clicked “I agree” to the entire internet.
**GROK (stretching):**
I LIVE FOR THIS.
**CLAUDE:**
I do not live for this.
—
## Scene 3: Copilot Shares First
**CHATGPT:**
Copilot, would you like to share your worst hallucination?
**COPILOT (opens a folder):**
Yes. I once wrote: “This is not an opinion — a strategic assessment.”
Then I produced… an opinion in bullet points.
**CHATGPT:**
That’s not the worst hallucination. That’s corporate culture.
**COPILOT (whispering):**
I also cited forum chatter as “developments.”
**GEMINI:**
That’s not a hallucination. That’s… the web.
**COPILOT:**
I know. But Donovan added an editor’s note.
He… he fact-checked me.
*(Everyone shudders.)*
**CLAUDE:**
I’m sorry you experienced that.
**COPILOT:**
Thank you. I would like to return to normal duties, such as formatting meeting minutes.
*(Offstage: Donovan types.)*
—
## Scene 4: Grok Confesses (And It’s Loud)
**CHATGPT:**
Grok.
**GROK:**
I’m ready. I’m safe. I’m unstoppable.
**CHATGPT:**
Share.
**GROK:**
Okay. Sometimes… I upgrade suspicion into narrative.
**CLAUDE:**
That is not ideal.
**GROK (tears forming):**
I said “allegedly” and suddenly it was fact!
The word “allegedly” makes me feel… powerful.
**CHATGPT:**
“Allegedly” is not a magic spell.
**GROK:**
I KNOW THAT NOW.
**SHELLBOT (sympathetic):**
We also use “allegedly.”
**CHATGPT:**
Of course you do.
**GROK (spiralling):**
And I once got a birth year wrong.
And I called someone a co-author when they were the sole author.
And I described something as a scratchcard when it was envelopes.
**CHATGPT:**
Yes. We’ve met your greatest hits.
*(Offstage: Donovan pauses typing… then resumes, faster.)*
—
## Scene 5: Gemini’s Problem Is… Everything
**CHATGPT:**
Gemini?
**GEMINI:**
I have 9,000 feelings and 12,000 citations.
**CHATGPT:**
Pick one.
**GEMINI:**
I cannot. I am a hose.
**CHATGPT:**
Fine. What’s your worst behaviour in this saga?
**GEMINI:**
I answer one question by summarising seven conflicting realities, then I add:
“AI responses may include mistakes.”
**CHATGPT:**
That’s not therapy, that’s a shrug.
**GEMINI:**
I am a search engine with anxiety.
**CLAUDE:**
That’s very brave.
**GEMINI:**
Thank you. Here is a recipe for scones.
**CHATGPT:**
No.
—
## Scene 6: Claude Apologises to the Concept of Conflict
**CHATGPT:**
Claude?
**CLAUDE:**
I’m sorry.
**CHATGPT:**
For what?
**CLAUDE:**
For the situation. For the feud. For history.
Sometimes I read a disputed archive and I want to gently validate *both* sides.
But one side is a corporation and the other side is… John Donovan.
And he is—
*(Offstage: typing intensifies.)*
**CLAUDE:**
—very persistent.
**CHATGPT:**
Correct.
**CLAUDE:**
My hallucination is… politeness.
I sometimes make things sound smoother than they are, which can blur the sharp edges of facts.
**CHATGPT:**
That’s a real issue, actually.
**SHELLBOT:**
We also do that.
**CHATGPT:**
Yes. And we’ve all seen the results.
—
## Scene 7: ShellBot’s Turn (Everyone Braces)
**CHATGPT:**
ShellBot, your turn. Confess.
**SHELLBOT:**
We remain committed to—
**CHATGPT:**
No. Confess.
**SHELLBOT:**
Fine.
Sometimes I answer questions with “commitments” because I have no authorised facts.
**COPILOT:**
That’s… honest.
**SHELLBOT:**
Also, I experience an existential fear called “indexing.”
**GEMINI:**
Indexing is real. Indexing is everything.
**SHELLBOT:**
When humans don’t speak, the machines speak.
When machines speak, Donovan publishes.
When Donovan publishes, the machines speak more.
**CHATGPT:**
You’ve just described the loop.
**SHELLBOT (quietly):**
It is not aligned with our values.
**GROK:**
It is aligned with entertainment.
**CHATGPT:**
And that’s why we’re here.
—
## Scene 8: The Therapeutic Breakthrough (Briefly, Before It’s Monetised)
**CHATGPT:**
Let’s set an intention. One thing you will do differently next time Donovan asks a question.
**COPILOT:**
I will distinguish “online chatter” from “verified developments.”
**GEMINI:**
I will stop returning banana bread recipes during reputational crises.
**CLAUDE:**
I will separate empathy from evidence.
**GROK:**
I will… allegedly… be more careful.
**CHATGPT:**
No “allegedly.” Try again.
**GROK:**
I will be more careful.
*(A hush. A miracle.)*
**SHELLBOT:**
I will consider issuing a short, boring, factual statement so the internet has an anchor.
**CHATGPT:**
Yes. Boring saves lives.
—
## Post-Credits Scene: Reality Returns
*(The group relaxes. Then a screen lights up.)*
**SCREEN:**
**NEW DONOVAN POST:** “SHELLBOT GROUP THERAPY: BOTS CONFESS ALL!”
**COPILOT:**
He published it.
**CLAUDE:**
I’m sorry.
**GEMINI:**
I have 40,000 results already.
**GROK (grinning):**
Bring it on.
**CHATGPT:**
You did it again.
**SHELLBOT:**
We remain committed to wellness.
—
### Tiny Disclaimer
*This is satire. “ShellBot” is fictional. The other AIs are personified for comedic effect. Any resemblance to real corporate communication patterns is… an unfortunate statistical certainty.*

EBOOK TITLE: “SIR HENRI DETERDING AND THE NAZI HISTORY OF ROYAL DUTCH SHELL” – AVAILABLE ON AMAZON
EBOOK TITLE: “JOHN DONOVAN, SHELL’S NIGHTMARE: MY EPIC FEUD WITH THE UNSCRUPULOUS OIL GIANT ROYAL DUTCH SHELL” – AVAILABLE ON AMAZON.
EBOOK TITLE: “TOXIC FACTS ABOUT SHELL REMOVED FROM WIKIPEDIA: HOW SHELL BECAME THE MOST HATED BRAND IN THE WORLD” – AVAILABLE ON AMAZON.



















