---
title: "Rocket Scientists Hooked Up ChatGPT to the Controls of a Spaceship, and the Results Were Not What You Might Expect"
description: "ChatGPT could perform surprisingly well when put in charge of navigating space inside a simulated spacecraft, researchers have found."
date: "2025-07-04"
modified: "2025-07-04"
authors:
  - name: "Victor Tangermann"
    job_title: "Senior Editor"
    link: "https://futurism.com/authors/victor"
url: "https://futurism.com/scientists-chatgpt-controls-spaceship"
categories:
  - "Artificial Intelligence"
  - "OpenAI"
tags:
  - "artificial intelligence"
  - "chatgpt"
  - "large language models"
  - "OpenAI"
  - "spacecraft"
---

# Rocket Scientists Hooked Up ChatGPT to the Controls of a Spaceship, and the Results Were Not What You Might Expect

![ChatGPT could perform surprisingly well when put in charge of navigating space inside a simulated spacecraft, researchers have found.](<https://futurism.com/wp-content/uploads/2025/07/scientists-chatgpt-controls-spaceship.jpg>)
*\<em\>Image: Getty / Futurism\</em\>*

## AI Kirk

Despite being [stumped by simple children's games](<https://futurism.com/frontier-ai-models-stumped-childrens-game>) and [uncontrollably hallucinating](<https://futurism.com/the-byte/openai-new-ai-problem-hallucinate-more>), AI models could perform surprisingly well when put in charge of navigating space inside a simulated spacecraft, researchers have found.

As [*Live Science* reports](<https://www.livescience.com/space/space-exploration/chatgpt-could-pilot-a-spacecraft-shockingly-well-early-tests-find>), a team of scientists from MIT and the Universidad Politécnica de Madrid, Spain, instructed OpenAI's blockbuster AI chatbot ChatGPT to "operate as an autonomous agent controlling a pursuit spacecraft."

To their amazement, as detailed in [a paper](<https://arxiv.org/abs/2505.19896>) slated to be published in the *Journal of Advances in Space Research*, they found that the large language model exceeded expectations, scoring second place in a space simulation competition based on the popular video game Kerbal Space Program that pitted several AI agents against each other.

## To Infinity and Beyond

To test how autonomous agents could be used to maneuver satellites and other space-based assets, researchers created a software design challenge called the [Kerbal Space Program Differential Game Challenge](<https://www.ll.mit.edu/conferences-events/2024/01/kerbal-space-program-differential-game-challenge>). Participants are encouraged to employ a number of different approaches, such as reinforcement learning.

As part of the program, scientists pitted various autonomous systems against each other in several different scenarios.

In the latest research, the team developed a way to translate instructions and queries about how to orient and maneuver a spacecraft into text and fed it to several commercially available LLMs. With only a small number of prompts, they found that ChatGPT, in particular, performed surprisingly well, coming in second place in the Game Challenge. First place, in an important caveat, went to a system based on equations that model spaceflight.

Best of all, the research was conducted before OpenAI released its GPT-4 model in March 2023, suggesting the latest iterations might perform even better.

But there are still plenty of reasons why we should think twice before handing over the controls to the tech. For one, the current crop of AI models still [suffers from rampant hallucinations](<https://futurism.com/ai-industry-problem-smarter-hallucinating>), which could make for an awkward — and particularly terrifying — twist in a deep space mission.

**More on AI:** *[Tech Workers Say They're Rapidly Being Replaced by AI](<https://futurism.com/the-byte/tech-workers-ai-lay-offs>)*

## Author
I've been at Futurism since 2017, where my role has evolved to encompass design, writing, and increasingly editing. I've always been fascinated by space exploration and advanced transportation, which I've leaned into by interviewing luminaries in those fields while closely following the dimensions of policy and regulation that allow next-generation projects to succeed -- or, sometimes, to fail. I'm also keenly interested in the effects of generative AI on society, policies, and democratic institutions, as well as clean energy, physics and biology, and the vagaries of tech leadership. My work for Futurism has been cited by publications including Ars Technica, Gizmodo, PC Magazine, Jalopnik, Fox News, and the New York Post. I spent my childhood living in locations including Manila, the Philippines, and Geneva, Switzerland, attended McGill University, and now live in Toronto, Canada. Before Futurism I worked at AskMen and a small photography studio. In my free time, I'm an avid gardener, foodie, and craft beer lover, as well as a maker of artisanal hot pepper sauces. I have a magnificent dog named Freida.

### Author social links  
[Bluesky](<https://bsky.app/profile/vtanger.bsky.social>)