Beyond speech and writing: could AI let humans exchange thoughts at high speed?

Modern computers exchange data at unimaginable speeds - a billion bits a second over an ordinary gigabit connection, and many times faster inside advanced data centres. Humans, by contrast, still share their thoughts almost entirely through speech, writing, gestures and body movements, and what matters in human communication is not the volume of information but its meaning.

A new scientific research essay and proposed research framework examines whether artificial intelligence could close that gap by reading the brain directly. There are two main approaches to recording brain activity, and recent studies have made striking progress: researchers including Littlejohn and colleagues processed signals from the human brain in 80-millisecond windows, achieving average communication rates of about 56 words per minute and mean accuracy of 81.25 per cent.

The proposed framework for "AI-assisted high-speed thought communication" would identify three key elements of a thought, capture them using non-invasive methods such as EEG, and use AI to convert the result into a form another person could understand. The essay argues the hypothesis is scientifically testable, and sets out how receiver comprehension could be measured and how the method's performance should be compared with typing, speech and other channels.

The authors acknowledge serious hurdles: non-invasive signals like EEG carry heavy noise and interference, human attention, memory and comprehension all shape the outcome, and any exchange of thoughts would have to be encrypted. Even so, they say, the approach could transform communication for people with disabilities.

Their conclusion: the goal is not merely to send more data faster, but to let meaning itself travel between minds.