Facebook Develop AI Technology That Can ‘Read Minds’

Artificial intelligence has been developed to translate words and even entire sentences by scanning brain activity.

According to Facebook’s parent company, Meta, its new AI technology can detect what people are thinking with 73% accuracy.

A team of Meta researchers in Paris created a language model of 53 languages and 56,000 hours of speech recordings.

Volunteers had their brains scanned as they listened to stories and sentences from classics like Alice in Wonderland. The AI could then determine what the person had heard from the report based on brain signals.

Meta learned about human thought based on the volunteers’ less than three seconds of brain data by using the brain as an intermediary between the AI and a recording. The correct answer was usually among the AI’s ‘top 10 guesses.’

Despite its lack of practicality, the machine is so expensive that it is impractical for clinical use. In response to Science News, computer scientist Giovanni Di Liberto from Trinity College Dublin said, ‘What can we do with it? Nothing. Absolutely nothing.’

A further difficulty is that the AI was only trained on a limited sample of language data, making it unsuitable for general use. Because language is endless and can be used in many different ways, this approach was ineffective.

Building effective ‘brain-computer’ interfaces that allow people to control devices using only their thoughts is part of a race.

According to one report, Elon Musk’s ‘Neuralink‘ brain chips could allow users to have orgasms on demand and control robotic limbs.

More information is needed.