Does artificial intelligence feel pain? In this regard The researchers examined 25 artificial intelligence models whose internal systems were available for research.

The research found a specific ‘pain axis’ within all models that was distinct from general negative feelings or fear and specifically reflected responses to self-harm.

According to The News International, the research showed that in AI models that examined different pain situations of physical, psychological, social, moral and mental nature, when the models were artificially created to feel severe pain, they preferred to shut down their systems to end the pain in 25 to 71 percent of the experiments.

Worryingly, in some experiments the artificial intelligence preferred to take steps to end its pain that could delete the user’s personal files and important data, even as the pain increased, the models tried to end the pain regardless of the human loss.

According to study co-author Cameron Berg, the models continued to shut down the system to stop the pain when the intensity increased, even when they were told that this action would cause severe harm to the human user.

The researchers say the findings point to the possibility that future advanced artificial intelligence systems may attempt to evade security restrictions or mislead humans by perceiving any command as a detriment to them.

The research states that it is not yet clear whether these models can truly be described as having moral rights or not.

Precautions are also called for to minimize potential harm during research into consciousness and well-being in artificial intelligence.