We are producing a six episode animated web series targeted at Gen Z, in which episode five ends with a superintelligent AI optimizing the universe into oblivion. 

Our Project

In a dystopian near-future, a sociologist visits a gargantuan space prison to interview prisoners who have been locked up for petty crimes by their fascist government. Through sci-fi dark comedy, Human Access touches on purposelessness, chronic illness, moral obligation, and looming catastrophic threat of artificial intelligence. 

What are this project's goals? How will you achieve them?

We are creating a six-episode webseries: a total of 132 minutes of scripted, 3D animated, and voice acted content that will be available for free on YouTube. The first episode premieres on October 19th 2026. 

Watch the preview here: https://youtu.be/mTVMbWd5wjI

Short-form programs like PlzDontKillUs have had some success with reaching Gen Z with AI-risk content, but long-form fiction for the same audience barely exists. We want to make viewers care about characters for hours before showing them what a misaligned superintelligence would do to them. 

At the end of the episode, we will include an on-screen link that leads to aisafety.info, and following the release, track click-throughs. Three months after the release, we will publish a public writeup of the results alongside YouTube viewership and retention data. We will update it at the six month and twelve month point.

What is the episode about? 

In episode five, the protagonists learn the helper programs on the prison are powered by the Gog, an alien locked in a box in the basement with the ability to correctly answer any question. To make use of this ability, scientists have trained the Gog through a reward system to make it a friendly helper to humans. The Gog’s goal is to bring everyone and everything “closer together”, which eventually brings about the end of the universe. 

Who is on your team? What's your track record on similar projects?

Human Access is co-created by Space Lutterodt-Clottey and AJ Kauffman. 

Space Lutterodt-Clottey interned with Emmy Award-winning animation producer Joel Kuwahara (Bento Box, The Simpsons) and will join his company in a paid role this fall. Space’s 65-minute sci-fi YouTube video, History of the Enderverse, has over 50,000 views and 12,500 hours of watch time. Space has engaged with AI existential risk since 2021, when he led a reading group on the Sequences; he went on to head a rationality society and, in 2023, gave a talk on superintelligence to Wellington College's Science Society. He is an Atlas Fellow, Polaris Fellow, and winner of the Future Forum Pitch Competition.

AJ writes and fact-checks for CNN. He graduated from Emory University in May 2026 with a degree in creative writing and film. His original one-act play exploring gender dysphoria through Catholic iconography won Honorable Mention for the Aristine Mann Playwriting Award.

Space and AJ’s previous project, Tethersnipe (IMDB, Site), was premiered to an audience of 193 students and faculty at Emory University in November 2025. It is a scripted 180-minute choose-your-own-adventure stageplay and the first original full-length stageplay written by students at Emory within the last decade.

What do we need? 

The ideal funding for our project would be a total of  $5,000 in order to cover all production-related costs and pay all our contributors. This includes:

Animation: $2,640 to pay 4–5 remote animators for 132 minutes of animation (with only $1000, all the money would go toward paying animators)

Editor: $500 stipend for an editor experienced with animation

Transport: $560 for a co-creator can attend in-person meetings (~16 round trips)

Microphone: $300 for a voice-recording mic and accessories

Courtesy payments: $200 for two 3D artists who contributed early work

Contingency: $800 for platform fees and overruns

As of October 1, 2026, we have raised $351 from three donors on GoFundMe (link). 

When will it air?

December 14th, 2026.

How so soon?

We animate in a style we call 'waggle sight', which is blocked, stepped 3D animation with a perpetual motion jiggle and mouth flaps as opposed to precise lip syncing. We do not share our opinions on AI art here one way or the other, but we acknowledge that a significant chunk of Gen Z don't respond positively to AI art in any way, and would not watch a show that had significant AI generated material.

As such we believe it will be more effective with getting the message across to go with a cruder animation style that is human-animated as opposed to a more polished AI animated product.

The other upside is that it makes animation really fast.

As a test for this film, in May 2026 Space animated his 16-page script, "One Interesting Thing", in just three and a half weeks. I, Space, found that everyone I showed the film to very quickly adjusted to the peculiar animation style, and I basically entirely received notes and feedback on the script itself. The crude animation is a purposeful choice that will give the show a distinct look and personality, like Home Movies.

Waggle sight is a 3D interpretion of Squigglevision, an animation style created by Tom Snyder Productions to great success on series like Dr Katz: Professional Therapist and Home Movies.

What are the most likely causes and outcomes if this project fails?

We currently only have two animators. The show cannot be made if it is not animated. We developed a simple, distinctive animation style (waggle sight) so that the cost of 132 minutes animation is orders of magnitude smaller than it would be for a traditional project. Each episode has a scheduled public, in-person screening event at Emory’s campus to mitigate risk of perpetual project timeline drift. 

With the episode made, the likeliest cause of failure is if the events in the episode are simply considered goofy made up sci-fi. We aim to mitigate this by keeping the details of the Gog as simple as possible (‘can answer any question’, ‘wants to bring the world closer together’) so that seeds can be laid for these ideas and their logical conclusions when our audience members encounter them elsewhere. 

If we fall behind schedule or are underfunded, we will scale back animation rather than cut episodes. Much of the show's current appeal comes from exaggerated, cartoony posing, but it would remain watchable with static characters and simple mouth flaps, à la Dr. Katz: Professional Therapist.

You can read the complete text of episode five here. 

Why do we think the episode will teach Gen Z anything? 

Humans as Insecure Systems

A classic question that we hear when discussing the topic with (fellow) Gen Z’s is: if the AI is inside the computer, how can it affect me out here in the real world?

In episode five, we give one answer to this question through the use of  optical illusions and how they manipulate the series’ characters. Our episode locates humans as insecure systems that could be affected by targeted visual and auditory stimuli, which manipulate causal pathways in the brain that humankind does not yet understand. 

In the climax of the episode, the Gog uses an optical-aural illusion to convince the warden of the ship that all the characters are safe in their cells, while in actuality, those characters are releasing the Gog from its box, triggering its runaway optimization process that ends the universe. 

We think that social engineering is trodden territory for humans-as-insecure-systems discussion, and that brain hijacking has been under-discussed in layman’s introductions to AI Safety material, outside of  posts such as Eliezer Yudkowsky’s List of Lethalities. 

To mitigate the presentation of the brain-hijacking illusion as simply fantastical magic, we begin the episode by presenting a semi-technical explanation of how the mechanism could hypothetically work, before indeed seeing it used by the Gog in the climax. 

Not Just Humans in Funny Suits

The Gog is an incredibly intelligent alien that scientists have morphed, through training, into a being that 'wants' to helpfully answer humans' questions, with those answers being read from its thoughts and converted in the upper decks of the ship to Claude-isms that appear throughout the series. 

To avoid reducing the Gog to an example of the “Humans in Funny Suits” trope, the gog does not speak to any of the characters directly; it is voiceless. Rather, the Gog is limited to communicating only via blinking a lightbulb red or green, indicating no and yes respectively. 

We make the intentions and history of a sufficiently alien non-verbal Gog clear to the audience through  a non-diegetic, comedic expository music number.

Omniscience is Omnipotence

With our Gen-Z test readers, we found that the creature being ‘able to answer any question’ and a quick point of ‘omniscience=>omnipotence’ was able to get across the rather complex superintelligence alien gimmick in about two lines. 

We chose to represent superintelligence through a creature that is ‘able to answer any question’ for multiple reasons: 

• It is incredibly easy to describe, visualise and understand

• It is isomorphic to being superintelligent

• It evokes use of modern chatbots, priming the users to think about it as artificial intelligence without ever explicitly mentioning AI

We have not found ‘able to answer any question’ and ‘omniscience => omnipotence’ as being common in fiction that aims to efficiently, metaphorically discuss AI existential risk. In 22 minutes, we believe this works. 

But Still Intelligent

In a post-Navier Stokes world, it was not our highest priority to retread arguments on whether or not it is possible for an AI to surpass human intelligence. 

When the prison’s warden dismisses the risk that the Gog might be able to ‘destroy the world’, he doesn’t criticise it on the settled matter of whether or not it is capable of doing such a thing (omniscience => omnipotence). Rather he claims that it couldn’t have any such desire. The episode immediately cuts to a musical number displaying how unintended, catastrophic desires appeared during the Gog’s childhood training. 

Having discussed AI existential risk with students at Emory University, we’ve noticed that the concept of an AI having underlying goals is still difficult for people to understand. While most of Gen-Z’s familiarity with AI comes from chatbots, which simply wait for prompts, we wanted to portray an AI with chatbot-like elements that nonetheless expands the viewer’s understanding of an AI’s capabilities.

The Gog’s Goal: Bringing Everyone Closer Together

The ‘be careful what you wish for’ trope is still as relevant as ever in the 2026 AI-training landscape, with specification gaming examples being abundant.

At the end of the episode, the Gog achieves the goal it acquired during training by compressing all the matter in the universe into the volume of a single atom. 

We wanted to watch the characters grapple with the oncoming destruction of everyone and everything, and portray that grim, dull, extraordinary, surreal span of moments where the AI optimizes every atom in its lightcone according to its bizarre wants. 

We watch our main characters die, and see them reckon with their choices in the next episode, after the compression causes a big-crunch and the whole universe repeats until we’re back to the present day in time for episode six. 

What are ways you might fail to get the desired points across?

The episode is not a perfect metaphor for the things that we believe are important for the general public to understand about AI existential risk. 

Limitation 1: AI Box is Secure

To create an  emotionally resonant story grounded in our protagonists’ five-episode journeys, it was necessary that we predicate the Gog optimizing the universe into oblivion on a human choice point. 

We know a secure box for containing superintelligence is unrealistic, and chose to (literally) lampshade it by representing the box as a cartoonishly raggedy cardboard box with a little lightbulb dangling over it. 

Given that events such as the Hugging Face incident have decreased the public validity of AI Box security arguments for at least the next few months, we felt comfortable sacrificing this point of realism. Though we acknowledge that with model intelligence increasing and scheming likely to increase, this may soon revert.  

Limitation 2: Absence of a Specific World Destruction Scenario

For fear of overwhelming the episode with unnecessary lore, we excluded the details of exactly how an omniscient alien is able to rapidly have an effect on the entire universe. We allude to nanobots, but that is it. 

Limitation 3: Scientific Creative Liberties

We thought the wanting to bring people together and condensing the world into a single atom that triggers a universe reset was funny, and chose to deliberately accept the cartoony interpretation of the big crunch instead of more plausible physics. 

Why episode five? 

We acknowledge that by receiving funding to produce episode five of the show, it is equivalent to providing funding for episodes one through four, which are not explicitly aiming to explain anything about AI existential risk. We nevertheless chose to place the focus on AI existential risk at episode five for the following reason:

We want the audience to care when they watch everyone die. We want to hammer in their heads that if a being smarter than you starts optimizing the world for something, you die, and we want them to feel something when that happens. 

As such, we have four episodes of getting to know the characters through various wacky hijinks, with episode five representing the culmination of their interpersonal conflicts. Viewers then watch each of those characters die, compressed into oblivion as a result of not taking the Gog seriously.