Well, long time no posting!
The post "A.I. trained to read minds and translate private thought to text via brain scans" from BoingBoing caught my eye. Here with extensive training on a specific person's brain activity while listening to spoken text, is able to correlate later brain activity (while watching silent films or thinking of speaking) and do pretty well at reconstructing at least some of what the person "had in mind". Note though that patterns for one person do not carry over to other people.
This language-decoding method had limitations, Dr. Huth and his colleagues noted. For one, fMRI scanners are bulky and expensive. Moreover, training the model is a long, tedious process, and to be effective it must be done on individuals. When the researchers tried to use a decoder trained on one person to read the brain activity of another, it failed, suggesting that every brain has unique ways of representing meaning.
