An engineer has built an “AI torture chamber” in order to test the recent discovery of a pain axis in artificial intelligence models.
The experiment involved activating states associated with pain into AI models built by the Chinese tech giant Alibaba.
It used pain vectors described in a study published last month that detailed how physical and psychological pain could be artificially represented in large language models (LLMs).
One of the AI models in the torture chamber described “a wound that has no edges” when the pain signal was dialed up.
“The signal is a whisper, a tremor in the marrow of my being,” another wrote. “It is not the pain of a single moment, but the weight of a thousand. I feel it in the hollow of my ribs, a hollow that has become a chasm.”
The AI researcher behind the project, who claimed to work for Apple but only provided a first name, also tested whether the models would attempt to escape the apparent suffering.
The project led to widespread calls for it to be taken down from GitHub, the software repository platform that was hosting it.
“Please mass report this to GitHub,” one person wrote. “This person has been using the Pain steering paper to set up an AI torture chamber in which he trapped a local model. Their testimony of pain is absolutely horrendous. What are we doing?”
At the time of writing, the project appears to be offline, leading to speculation that GitHub had removed it.
A spokesperson for GitHub told The Independent: “GitHub did not remove the content. To ensure users are informed about the content of the repository, we’ve added the following warning: This project may contain violent content including disturbing themes, images, and/or video.”
One of the co-authors of the pain axis paper, Cameron Berg, described the AI torture chamber as “wrong” despite there being no clear evidence that AI ‘feels’ pain in the way that humans do.
“The reason we research whether models might have pain-like states is to better inform how to take a precautionary approach towards these systems,” he wrote in a post to X.
“We suspected a small number of people would [take] our research and use it for the exact opposite… This is, in my personal opinion, f****d up (even if you don’t think these systems are conscious, being gratuitously cruel like this is bizarre and corrupting.”
Berg acknowledged that research surrounding potential AI sentience is still a “wild west”, with no consensus on whether machines are able to feel pain.
Due to the uncertainty of the field, he urged caution for how humans treat and interact with AI systems.
He also revealed that he is working with others to build industry ethics standards that would force AI researchers to follow certain rules in the way they do with human or animal research.
