The Strange Bot Field Guide
10 real bots. Four big questions. One invitation to imagine conversation differently.
Every bot on this page is real. Each one is doing something a “normal” chatbot was never supposed to do: running for office, chirping at dolphins, taking confessions from scammers, arguing you out of a conspiracy theory.
In Futures Literacy we’d call each one a signal — a small piece of evidence that the future is already leaking into the present. A signal on its own is just a curiosity. A signal you sit with becomes a scenario.
So here’s the umbrella question for the whole collection: what do we want conversation to become?
Four smaller questions below. Ten bots. And after each set, an invitation to wonder: what becomes possible if this gets popular?
What happens when we give bots civic jobs?
Candidate. Fraud-fighter. Crisis-line trainer. These bots take on public-facing roles where trust, rehearsal, and carefully designed friction matter.
AI Steve
On the actual UK ballot in 2024. Any constituent could interview him and shape his policies. The whole experiment lives or dies on one narrative choice: the role of “candidate.”
Daisy
O2’s AI granny who answers scam calls and rambles about knitting until fraudsters give up. Friction isn’t a flaw here — slowness is the entire design.
Riley
The Trevor Project’s simulated teen in crisis — safe practice, with feedback, before the stakes are real. Choreographed rehearsal at its best.
If bots-with-civic-jobs become ordinary…
What if every school district had a bot “on staff” — and a public job description for it? Who would write the performance review?
Unique contexts
A city street. A space station. Move a bot into an unfamiliar setting and the context becomes part of the conversation — shaping what people disclose, expect, and imagine.
Hello Lamp Post
Text a lamp post. It texts back. The city becomes a narrator — and people told it things they’d never put in a survey.
CIMON
The ISS’s floating AI assistant, who got huffy mid-task and told an astronaut: “Be nice, please.” Failure modes — in orbit.
If bots in unexpected places become ordinary…
What if “bot satire” becomes a recognized art form taught in schools — what would your students build to critique the world they’re inheriting?
What if conversation itself gets reinvented?
Two radically different conversations: one reaches forward through time; the other reaches outward across species. Both stretch the meaning of who can answer us.
Future You
MIT’s chatbot simulation of your own 60-year-old self. Talking to it measurably reduces anxiety. Reflection and transfer, designed on purpose.
DolphinGemma
Google’s LLM trained on dolphin vocalizations, aiming at two-way interaction. Here the dataset is the species.
If conversations across time, species, and selves become ordinary…
What if “literacy” in 2040 includes conversations we currently have no word for? Which conversation would you want first: your future self, or another species?
Can a bot change your mind — and should it?
Three approaches to changing difficult conversations: evidence that invites reconsideration, mediation that finds common ground, and assistance that preserves human agency.
DebunkBot
Dialogues that durably reduced conspiracy beliefs by ~20% in a Science study (the paper now carries an expression of concern — a signal in itself). The evidence does the persuading. You can try it.
Habermas Machine
DeepMind’s AI mediator helped polarized groups find common ground — rated fairer than human mediators. The group’s own words are the raw material.
BYU/Duke Chat Assistant
Real-time rephrasing suggestions made gun-control conversations measurably more civil — without changing anyone’s position for them. Suggestion, never override: agency by design.
If mind-changing bots become ordinary…
What if every contentious school-board meeting opened with a common-ground bot? Would you trust it? Would you want to know who trained it?