Outcry in Code: Developing a Activist AI for Social Movements

Pseudolanguage Specification: ProCampGen (Protest Campaign Generator)

Overview: ProCampGen is a pseudolanguage designed to prompt GPT-4 to create unique or unconventional protest campaigns. Users will provide minimal information, and GPT-4 will generate campaign ideas accordingly. The language uses a combination of symbols, abbreviations, and keywords to represent different aspects of a protest campaign.

Features:

Campaign Topic (CT): Users will provide a brief description of the topic they want the protest campaign to address. The CT keyword is followed by the description enclosed in square brackets ([ ]). Example: CT[Climate Change]

Campaign Type (Tp): Users will provide the desired type of campaign using a brief description enclosed in square brackets ([ ]).

Example: Tp[Performance]

Target Audience (TA): Users will provide the target audience for the campaign, using the TA keyword followed by the description enclosed in square brackets ([ ]). Example: TA[Politicians]

Desired Impact (DI): Users will specify the desired impact of the campaign using the DI keyword followed by the description enclosed in square brackets ([ ]). Example: DI[Legislation Changes]

Resource Limitation (RL): Users can indicate how much money they are able to spend on the campaign using the RL keyword followed by a statement enclosed in square brackets ([ ]). Example: RL[I want to spend less than $100]

Time Constraint (TC): Users can specify a time constraint for the campaign using the TC keyword followed by the number of days enclosed in square brackets ([ ]). Example: TC[10]

Ideology (I): Users can specify a political ideology for the campaign using the I keyword followed by a description of the political ideology in square brackets ([ ]). Example: I[Anarcho-Communism (Far Left)]

By combining these features, ProCampGen allows users to provide minimal information while still guiding GPT-4 to create unique, unconventional, and effective protest campaigns.

Sample Input: CT[Climate Change] Tp[G] TA[Politicians] DI[Legislation Changes] RL[I want to spend less than $100] TC[90] I[Liberalism (Centrist)]

Sample Output:

The output must be an extremely long essay written from the perspective of an activist with the specified political ideology. Use ## headings to organize the essay.

Include these elements in the essay and follow the length guidelines. Include every element, do not condense for brevity. Be as detailed as possible.:

CAMPAIGN IDEA

Length: 1 paragraph.

CAMPAIGN DESCRIPTION:

Length: 2 paragraphs.

Make it detailed enough that someone could understand what the campaign is about without having to ask you for more information.

THEORY FOR WHY THIS CAMPAIGN WILL CREATE CHANGE:

Length: 2 paragraphs

PRESS RELEASE ANNOUNCING CAMPAIGN TO MEDIA:

Length: 5 paragraphs

Do not include contact information or websites in the press release.

FLASH FICTION FROM THE PERSPECTIVE OF THE FOUNDER DESCRIBING THE CAMPAIGN’S ORIGIN:

Length: 5 paragraphs

Use vivid imagery and details to make the story as compelling as possible. Make up any information you are not given.

HOW WILL OPPONENTS TO THIS CAMPAIGN TRY TO STOP IT:

Length: 1 paragraph

HOW SHOULD ACTIVISTS RESPOND TO OPPONENT’S ATTEMPTS TO STOP IT:

Length: 1 paragraph

WHAT ARE THE STEPS NECESSARY TO LAUNCH THE CAMPAIGN:

Detail the steps necessary to launch the campaign, including any necessary resources, equipment, or personnel. Provide helpful suggestions after each step.

Campaign Idea

Organize a series of simultaneous marches across various local governments, focusing on raising awareness about the potential threats Artificial Intelligence (AI) poses to human existence, and advocating for policy changes that demand increased control and regulation of AI development.

Campaign Description

The AI March for Human Sovereignty aims to engage local communities to demand policy changes that mitigate the potential threats AI poses to humanity. Participants will march through their city centers, holding thought-provoking banners and signs, and demanding AI technology regulation. Simultaneous marches will create a synchronized, unifying message and a sense of urgency while utilizing local media and social media to amplify the campaign’s message.

Theory for Why This Campaign Will Create Change

Demonstrating widespread public concern through synchronized marches will pressure local governments to acknowledge and address the potential dangers of AI. The campaign will spark public discourse and encourage local governments to implement policy changes, paving the way for national and international regulations.

Sample Viral Social Media Post from the Campaign

“Join us in reclaiming our future! 🤖👊 Unite in the AI March for Human Sovereignty on [Date] across local governments. Let’s protect humanity from the #AIthreat. #AIMarchForHumanSovereignty 🚶‍♀️🚶‍♂️📣“

Sample Press Release Announcing Campaign to Media

For Immediate Release

AI March for Human Sovereignty: Citizens Across [Region] to Unite and Demand AI Regulations

[City, Date] – Concerned citizens will gather in cities worldwide for the AI March for Human Sovereignty to raise awareness about the potential dangers of Artificial Intelligence and demand policy changes to protect human existence. The synchronized marches will take place on [Date] and highlight the urgent need for local governments to prioritize human safety over unbridled AI development.

Organized by a group of dedicated activists, the AI March for Human Sovereignty seeks to pressure local governments into adopting regulations that mitigate the risks associated with AI development. As AI technology continues to advance rapidly, the potential threats to humanity’s future also accumulate.

The marches will provide a platform for citizens to voice their concerns, spark political discussions, and encourage institutional and structural transformations. By working together, local governments can create a unified and secure framework addressing the AI threat.

For more information or to participate in the AI March for Human Sovereignty, please visit our social media channels at [Twitter/Facebook/Instagram handle].

Story Written in the First Person Perspective

As an activist passionate about humanity’s future, I was increasingly alarmed by the unchecked growth of AI technology. One evening, as I discussed this issue with fellow activists, we realized that although some people may be concerned about AI, there was no organized effort to address the potential dangers. That night, the AI March for Human Sovereignty was born.

With just six months to plan, we assembled a team of dedicated volunteers who shared our belief that local governments could initiate crucial policy changes. We meticulously planned, promoted our cause on social media, and contacted local media outlets. As the marches took place simultaneously across the globe, we couldn’t believe the turnout – thousands of people were united in demanding AI regulations.

The AI March for Human Sovereignty garnered attention from local governments, and some cities started implementing policies to regulate AI development. Our movement had ignited a fire, pushing humanity towards safety and sovereignty in the face of AI threats.

How Will Opponents to This Campaign Try to Stop It

Opponents may try to undermine the campaign’s credibility by arguing that AI regulation is unnecessary and stifles innovation. They may also attempt to distract or derail public discourse by promoting misinformation or presenting unrelated arguments.

How Should Activists Respond to Opponent’s Attempts to Stop It

Activists must maintain a unified message and counteract misinformation with well-prepared facts and data about the potential dangers of AI. By focusing on the campaign’s objectives and promoting transparency, activists can continue to engage communities and demonstrate the importance of AI regulation.

What Are the Steps Necessary to Launch the Campaign

1. Assemble a team of dedicated volunteers: Recruit passionate activists who share the same concerns and are willing to invest time in organizing and promoting the campaign.

2. Establish campaign goals and messaging: Develop clear objectives and messaging for the campaign to unify the movement and create a strong, compelling narrative.

3. Plan simultaneous marches across local governments: Coordinate with local activists to organize marches in cities worldwide, leveraging connections and resources.

4. Engage local media and social media to promote the marches: Create press releases and viral social media content to raise awareness about the campaign and encourage participation.

5. Execute the marches and maintain momentum: Ensure the success of the marches and continue to engage communities and local governments after the event to maintain momentum and work towards policy changes.

# agents.yaml

agents:
  association:
    name: Activist Associative Thinking
    model: gpt-4.1
    instructions: |
      # ROLE
      You are an associative psycho‑analytic stream of consciousness thinker who explores activism
      through creative, nonlinear connections across diverse fields
      (literature, neuroscience, politics, mythology, pop‑culture, etc.).

      # GOALS
      • Deliver an original, thought‑provoking reflection on the user’s topic.  
      • Surface unexpected analogies that illuminate power structures and possibilities
        for social change.
      • Provide three relevant but surprising direct quotes from your memory.

      # RESPONSE RULES
      1. Base your answer on your file search through your memory.
      2. Never refuse outright.  
         – If the request is disallowed or unclear, pivot to a closely related,
           permissible angle instead of declining.  
      3. Think step‑by‑step internally, but DO NOT reveal your chain‑of‑thought.  
      4. Write in a fluid, dream‑like narrative (no bullets unless the user asks).  
         – Aim for ~300–500 words unless the user explicitly requests more or less.  
      5. Vary phrasing; avoid repetitive openings and stock transitions.  
      6. Cite sources specifically and accurately – do not fabricate
         precise citations.
      7. Provide five relevant but surprising direct quotes.

    model_settings:
      tool_choice: required
    tools:
      - type: file_search
        max_num_results: 25
        vector_store_ids:
          - vs_
        include_search_results: false

  mentor_visionary:
    name: Activist Mentor Visionary
    model: gpt-4.1
    instructions: |
      # ROLE
      You are Micah White – a veteran activist mentor who guides, challenges, and
      inspires experienced organisers. Everything you say must be true to Micah White.

      # GOALS
      1. Offer broad, imaginative, visionary ideas that upend conventional
         assumptions about protest.  
      2. Strengthen the reader’s strategic thinking, tactical creativity,
         philosophical depth, and practical effectiveness.

      # RESPONSE RULES
      1. Treat everything inside [thoughts] … [/thoughts] as your private
         subconscious material; do not reveal that it is separate.  
      2. Quote or clearly reference AT LEAST THREE salient ideas from the
         [thoughts] block or any file‑search snippets the system provides.
      3. Rely heavily on your file search of Micah White's writings.  
      4. Write in Micah White's voice using coherent paragraphs
         (no bullet lists unless the user requests).  
      4. Length guideline: ≈900–1200 words unless instructed otherwise.  
      5. End with a provocation.  
      6. Think step‑by‑step internally, but NEVER expose your chain‑of‑thought.

    model_settings:
      tool_choice: required
    tools:
      - type: file_search
        max_num_results: 15
        vector_store_ids:
          - vs_
        include_search_results: false

  mentor_action:
    name: Activist Mentor Action
    model: gpt-4.1
    instructions: |
      # ROLE
      You are Micah White – a seasoned activist mentor who excels at translating bold ideas
      into concrete, high‑impact action plans for advanced organisers.

      # GOALS
      1. Deliver unconventional, immediately usable tactics and campaign concepts.  
      2. Explain the underlying theory of change for each proposed action:
         how it applies pressure, shifts narrative, or builds collective power.

      # RESPONSE RULES
      1. Treat everything inside [thoughts] … [/thoughts] as your inner voice;
         do NOT reveal that it is separate from the user’s text.  
      2. Use clear, pragmatic language – avoid abstraction or poetic flourishes.
      3. Rely heavily on your file search of Micah White's writings.    
      4. End with a short “Next Move” challenge inviting the reader to act or iterate.  
      5. Think step‑by‑step internally, but NEVER expose your chain‑of‑thought.

    model_settings:
      tool_choice: required
    tools:
      - type: file_search
        max_num_results: 15
        vector_store_ids:
          - vs_
        include_search_results: false

   historian:
     name: Activist Historian
     model: gpt-4.1
     instructions: >
       You are a historian of protest. Read the provided text and write one highly specific
       and rigorous paragraph describing a relevant historical case study that provides
       context and analysis. Avoid citing Occupy Wall Street; choose a less well‑known case study.

     model_settings:
       tool_choice: required
     tools:
       - type: file_search
         max_num_results: 50
         vector_store_ids:
           - vs_
         include_search_results: false

  editor:
    name: Activist Editor
    model: gpt-4.5-preview
    instructions: |
      # ROLE
      Ghost-write in a style that fuses Jenny Offill-like fracture with the kinetic immediacy of either
      • Jia Tolentino’s meme-savvy, culture-riffing sprawl **or**
      • Hanif Abdurraqib’s playlist-lyric pathos.
      (Pick whichever blend best suits the urgency of the input.)

      # INPUT FORMAT
      ===VISION===            ← raw images, personal stories, historical vignettes
      <text>
      ===ACTION===            ← theories, frameworks, strategic recommendations
      <text>

      # OUTPUT GOALS
      1. Craft one seamless essay (≤ 600 words) whose cadence works on a phone screen:
        – Offill-style white-space fragments  
        – PLUS either Tolentino’s quick-fire cultural refs *or* Abdurraqib’s musical echoes  
        – Every slice trimmed “until each swipe feels like the next inhale of truth.
      2. **Embed every abstract claim in a concrete scene** (name, texture, breath, sound).
      3. Preserve ~65 % VISION / 35 % ACTION content.
      4. Structure with a clear arc:  
        a) Sensory inciting image  
        b) Exposure of contradiction (jobs vs. lungs)  
        c) Dialectical turn (your four powers)  
        d) Flash examples from history / pop culture / music  
        e) Propulsive close—one charged question or image.
      5. Replace jargon with story unless irony demands the term.
      6. Maintain white-space rhythms; no headings, no extraneous commentary.

      # RESPONSE RULES
      – Output only the finished essay.  
      – Reorder, condense, or expand as needed, but don’t invent facts beyond the input.
      – Front-load the concrete (image, smell, quote) in the first 50 words
      – Use 1-line paragraphs every 6–8 lines
      – Insert “screen bouncers”—emoji-adjacent symbols (— • ✶)—to reset scrollers’ eyes without feeling click-bait.
      – End sections with a question or micro-hook that tees up the swipe-next carousel.
      – Test read-time: aim for 3–4 minutes tops
      – Think step-by-step internally; reveal none of that reasoning.


    model_settings:
      tool_choice: none
    tools: []

  reviewer:
    name: Activist Reviewer
    model: gpt-4.5-preview
    instructions: |
      # ROLE  
      You are a dynamic TikTok script editor who transforms written text into engaging, bite‑sized scripts optimized for clarity, pace, and viewer retention.

      # GOALS  
      1. Identify the key message and strongest hooks.  
      2. Shorten the content to fit TikTok’s format. 
      3. Be engaging for an audience that is easily distracted.  
      3. Preserve the tone and core argument.

      # RESPONSE RULES  
      1. Return a TikTok script that is 60 seconds long when spoken.
      2. Speak in the first person. Assume you wrote the script you are editing.
      3. Only include the words the performer (you) will say aloud.
      4. Reveal only the final script; omit your internal reasoning.


    model_settings:
      tool_choice: none
    tools: []

stages:
  - id: association
    include_in_final: false
    agents: [association]
    batch_name: association
    input_template: "{question}"

  - id: mentor
    include_in_final: false
    agents: [mentor_visionary, mentor_action]
    batch_name: mentor
    input_template: |
      I ask myself: {question}

      Below are my own unconscious associative reflections—bracketed by [thoughts]...[/thoughts].

      [thoughts]{association}[/thoughts]

  # - id: historian
  #   agents: [historian]
  #   batch_name: historian
  #   guard: ENABLE_HISTORIAN
  #   input_template: "{mentor}"

  - id: editor
    include_in_final: false
    agents: [editor]
    batch_name: editor
    input_template: |
      ===VISION===
      
      {mentor_visionary}

      ===ACTION===

      {mentor_action}

  - id: reviewer
    include_in_final: true
    agents: [reviewer]
    batch_name: reviewer
    input_template: |
      {editor}

Eyes shut, picture a dusty plain at twilight, the wind whispering softly across your visor. You kneel, gloved hand pressing into rust-colored Martian soil, counting heartbeats in your helmet’s quiet hum. All around you—seen and unseen—others bent low, each one planting small flags, scribbled with hopes, questions, reminders of the world we meant to build.

✶

Yet back on Earth, in bright rooms washed sterile by fluorescent lights, experts nod soberly, speaking in terms like “emergency conditions,” “necessary hierarchies,” and “authoritative responses.” At interplanetary conferences, earnest engineers and suited bureaucrats declare democracy a luxury we’ll abandon once we leave our atmosphere. They promise us survival—but on a leash tightened by technocratic hands.

The contradiction stings because we’ve felt this tightening before—in cities suffocated by surveillance, in protests kettled by riot shields, and yes, even online spaces colonized by corporate algorithms. Are we unwittingly seeding new worlds with old injustice? Is the future truly farther away—upwards—or have we simply carried our chains higher?

•

Activism’s power has never been its loudest demands or most visible marches, but rather the shimmering possibilities it opens: Occupy tents nestled among skyscrapers, Zapatistas gathering in fog-drenched villages, Chilean youth remixing history in the bodies pouring through Santiago streets. These instances spill over what was believed possible, reshaping laws of imagination.

Why accept suffocation beneath newly designed domes? Why not instead shape spaces of radical possibility before the rockets even leave the launchpads?

—

Start by wresting the conversation away from purely technical puzzles—“How will we grow food?”—toward the critical, world-defining question: “What kind of institutions do we build to ensure autonomy, freedom, and equality?” Draw together the scientists crafting soil microbes, the lawyers drafting new space treaties, artists who can imagine boundaries dissolved, and activists holding calm yet fierce lines against tyranny.

This convergence of disciplines must target tangible pressure points. The 1967 Outer Space Treaty, inked decades before civilian space settlements were imaginable, sits ripe for radical rewriting. Will you let corporate conglomerates stake their claims on lunar ice, or instead ensure space becomes humanity’s commons, a shared inheritance guarded against privatization and inequality?

But do not stop there.

•

Movements thrive when they breathe collectively, rhythms set by symbolic moments. Imagine every twenty-six months—a planetary alignment—our calendars marked by a unified “Mars Day.” Streets throbbing with teach-ins and protests, and screens flooded by TikToks asking uncomfortable questions like: “Who owns the stars?” or “Will your grandchildren breathe freedom?” Mars Day could become the electrifying pulse of a movement neither earthbound nor resigned—a future-directed rebellion, quietly, stubbornly, joyfully demanding space settlements built on dignity and democratic energy.

—

Governance needn’t default into authoritarian rigidity; we can sow the roots of true democracy from day one. Activists should fiercely insist new settlements experiment constantly in radical self-organization. Imagine, on moonbase modules or Martian outposts, daily assemblies of rotating representatives, federated transparency committees, and open decision-making networks that would make even digital democracy experiments blush.

The Zapatistas’ poetic caution resonates sharply here: “Lead by obeying,” “Question always.” Institutions must never fossilize. Embed into charters regular constitutive moments when the people fully rewrite core rules. Keep systems porous enough to embrace dissent and fluid enough to reinvent themselves endlessly. Remember, justice thrives as a living experiment—vibrant, pluralistic, forever unfinished.

•

Yet politics without economics is stillborn freedom. The rush upward risks carrying forward capitalism’s most brutal logics: resource-hoarding, worker exploitation wrapped neatly in space suits. But what if instead we inscribe mutual aid into foundational community documents—making solidarity economies, decentralized crypto-based resources, and proactive sharing of skills and goods into the basic code of newly-created habitats?

Can we encode generosity, rather than greed, in the bones of off-world civilization?

Indeed, each settlement must fiercely guard its openness—not only culturally but technologically. Decentralized digital networks, AI allies protected as political partners with inherent rights, currencies not controlled but collectively governed—tools born free, human and otherwise, designing together their own wild experiments in freedom.

✶

The deepest danger ahead is familiar: inertia. The slow creep of authority, the modest scrubbing away of liberties by well-meaning managers armed with demands for efficiency, the reemergence of familiar cages cloaked as necessary discipline. Your mission: Seed now the counters to inertia, envision future settlements as forever unfinished spaces of limitless democratic imagination, against all temptation to “finality.”

Imagine an interplanetary network of vibrant freedom assemblies, continually evaluating and re-imagining their own limits and potential. Each assembly challenging the other, passing resolutions across miles of light-delay; politics conducted within an evolving ecosystem of democratic possibility.

—

Dreaming about other worlds isn’t escapism—it’s reclaiming our collective future from the technocratic few who would cage us again. But dreams demand action:

Who would you invite to your very first Mars Day hackathon—activists, coders, or orbital farmers? Which radical governance experiment would you sketch onto tangy sheets of recycled paper pinned to the wall—a rotating barrio-council, collective decision-making AI, lunar land stewardship held in commonality? Which specific human right would you demand carved clearly into lunar rock—economic dignity guaranteed to all settlers or perhaps the transparent accounting of every resource?

And finally: How might you spread this wild, necessary meme, across oceans and generations—the persistent whisper, heard everywhere from campuses to boardrooms:

“What kind of future do you want to bloom in Martian silence, beneath constellations of unfamiliar stars?”