Scientists have developed an AI language decoder that can translate human thoughts into text without the need for any invasive procedures. This breakthrough technology uses an artificial intelligence transformer similar to ChatGPT and works by interpreting the gist of stories that human subjects watch, listen to, or imagine. The decoder can also predict the words that the user was hearing or saying or imagining.
What is the GPT AI language decoder?
The language decoder is an AI transformer that can interpret the gist of stories that human subjects watched or listened to or simply imagined. This achievement allows it to read peoples’ minds with unprecedented efficacy. The GPT AI language decoder can also predict the words that the user was hearing or saying or imagining.
Unlike existing techniques that use invasive implanted electrodes, the GPT AI language decoder works by focusing on the flow of blood through the brain, which is what is captured in fMRI machines. It works at the level of ideas, of semantics, and of meaning, allowing the team to capture the gist of the stories they analyzed.
How was it created?
The decoder was created with the help of three human participants who spent 16 hours in an fMRI machine listening to stories. The researchers trained an AI model referred to as GPT-1 in the study on Reddit comments and autobiographical stories. This way, it could learn which words and phrases were associated with certain brain patterns. Once that phase of the experiment was complete, the participants had their brains scanned in an fMRI while they listened to new stories that were not part of the training dataset. The decoder was able to translate the audio narratives into text as the participants heard them, though these interpretations often used different semantic constructions from the original recordings.
Why it matters
The potential applications of this technology are vast, including helping people with neurological conditions that affect speech to clearly communicate with the outside world. However, the researchers who created the decoder also warn about the potential for misuse of the technology for surveillance by governments and employers. Therefore, it is essential to ensure that the development of brain-reading platforms respects mental privacy. While the decoder is not yet practical for patients with speech conditions, researchers hope that future iterations of the device could be adapted to more convenient platforms.
