The system developed at the University of Texas in Austin (USA) - an artificial intelligence connected to an MRI scanner read the thoughts of volunteers. For now, it still makes mistakes, but in the future, similar devices could help people who are unable to communicate normally.
During the experiment, volunteers listened to a story or imagined telling it. Their brains were observed with the help of magnetic resonance imaging (MRI), and the artificial intelligence connected to it turned people's thoughts into text consistent with them. Importantly, no electrodes were needed to be implanted in the brain, and the person using the invention was not limited to a predetermined list of words, the researchers emphasize. However, the system requires special training - each person using it listens to podcasts for several hours, during which the computer watches his brain. "For non-invasive methods, this is a real leap forward compared to what was previously achieved, i.e. usually reading single words or short sentences" - says Prof. Alex Huth, author of the paper, which appeared in the journal "Nature Neuroscience". "Our model decodes long-term, continuous speech on complex topics," he emphasizes.
For now, the computer is not very accurate - it reads thoughts quite precisely in about 50 percent. However, he can often convey the meaning of the utterance. For example, he translated the thought, "I don't have a driving license yet," into "she hasn't started learning to drive yet." Listening to the thought, "I didn't know whether to scream, cry, or run away," he read as "She started screaming and crying and then she said 'I told you to leave me." The system was also good at reading the minds of the volunteers when they watched the videos. The creators of the program also addressed the topic of its possible abuses. They assure that at least currently it is impossible to read someone's mind if the person does not want to. It is also impossible to do it with someone with whom the system has not undergone a long training. “We take concerns about abuse very seriously and are working to prevent them. We want to make sure that people use these technologies when they want to, and that it helps them," said Jerry Tang, who led the research. Scientists hope that their idea will allow for the construction of devices that will allow people currently unable to communicate with the world, e.g. after severe strokes. The current version of the system only allows for its use in the laboratory, but according to the researchers, this can be changed - for example, instead of a large MRI scanner, it is likely that a much smaller device designed for non-invasive functional near-infrared spectroscopy can also be used.
#austin #usa #ai #artificialintelligence #technology #tech #future #science
After three years, the German coronavirus warning app stops warning users about exposure to infection. It is not even known how effective this application was.
According to former German chancellor Angela Merkel, the application was supposed to be a milestone in the fight against coronavirus.
Three years have passed, all pandemic restrictions have been lifted, and the most important feature of the application has just been disabled: warnings after contact with a person infected with Covid-19.
"It makes no sense (to keep alerts) with the low incidence we have now," said German Health Minister Karl Lauterbach. He added that Covid-19 is no longer so severe also due to the high immunity of the population.
On June 1, the coronavirus alert app will go into sleep mode. This means it will no longer be updated and will also disappear from the Google and Apple app stores. However, users can keep the app on their mobile phone if, for example, they have saved their vaccination certificates there and want to continue using them. Lauterbach also urged users to keep the app in case of another Covid-19 outbreak - or even a pandemic. - It may be that we need to reuse it for Covid-19. But it may also happen that we develop it further for other infectious diseases,' he said.
According to the German Ministry of Health, since its launch almost three years ago, the app has been downloaded a total of 48 million times. However, it is impossible to say how many people actively used it. Obtaining accurate statistics is not possible because the data from the application was only stored locally on the mobile phone for data protection reasons.
The German Ministry of Health is to investigate how effective the application was. The app cost the government €220 million – much more than originally planned.
Copyright