Researchers at the University of Texas at Austin have developed an artificial intelligence (AI) system capable of interpreting and reconstructing human thoughts.
The scientists recently published an article in Nature Neuroscience in which they examined the use of AI to non-invasively translate human thoughts into words in real time.
According to the researchers, current methods for decoding thoughts into words are either invasive — meaning they require surgical implantation — or limited in that they “can only identify stimuli from a small set of words or phrases.”
The Austin team circumvented these limitations by training a neural network to decode functional magnetic resonance imaging (fMRI) signals from multiple areas of the human brain simultaneously.
In conducting this experiment, the researchers had several test subjects listen to podcasts for hours while an fMRI machine non-invasively recorded their brain activity. The resulting data was then used to train the system on a specific user’s thought patterns.
After the workout, subjects had their brain activity monitored again while listening to podcasts, watching short films, and silently imagining telling a story. During this part of the experiment, the AI system was fed the subjects’ fMRI data and decoded the signals in real time into plain text.
According to a press release from the University of Texas at Austin, the AI was able to get things right about 50% of the time. However, the results aren’t exact – the researchers designed the AI to convey the general ideas being thought about, not the exact words being thought about.
Luckily for anyone concerned about having their minds infiltrated by AI against their will, scientists are well aware that this is currently not possible.
The system only works when trained on a specific user’s brainwaves. This makes it useless for scanning people who have not spent hours providing fMRI data. And even if such data were generated without a user’s permission, the team ultimately concludes that both decoding the data and the machine’s ability to monitor thoughts in real time require active participation on the part of the scanned person.
However, the researchers noted that this may not always be the case:
“[O]Our privacy analysis suggests that subject cooperation is currently required for both training and using the decoder. However, future developments could allow decoders to circumvent these requirements. In addition, decoder predictions, even if inaccurate without subject involvement, could be intentionally misinterpreted for malicious purposes.”
In related news, a team of researchers in Saudi Arabia recently developed a method to improve precision in diagnosing brain tumors by processing MRI scans through a blockchain-based neural network.
In their paper, the Saudi researchers show how processing cancer research on a secure, decentralized blockchain can improve precision and reduce human error.
Related: What is immutable, explained
While both of the above experiments are cited as early work in their respective research papers, it is worth noting that the technology used in each is widespread.
The AI underlining the experiments conducted by the University of Texas at Austin team is a generative pre-trained transformer (GPT), the same technology on which ChatGPT, Bard and similar large language models are built.
And the Saudi Arabian team’s cancer research was conducted using AI trained on Nvidia GTX 1080s, GPUs available since 2016.
Realistically, nothing will stop a clever developer (with access to an fMRI machine) from combining the two ideas to create an AI system that can read a person’s mind and record it on the blockchain.
This could lead to a “proof-of-thought” paradigm, where people might be able to mint non-fungible tokens (NFTs) of their thoughts, or record immutable books of their feelings and ideas for posterity, legal purposes, or just with rights.
For example, the impact of NFT coining thoughts on blockchain could have implications for writing texts and patent filings that use blockchain as evidence of exactly when a thought or idea was recorded. It could also allow prominent thinkers such as Nobel Prize winners or contemporary philosophers to codify their ideas into an immutable record – one that could be marketed and serve as a collectible digital asset.
#Scientists #Texas #developed #GPTlike #system #read #minds
