AI can “read your mind.”

AI

 

AI can “read your mind.” This is from a podcast by Dr. Phil Stieg, an esteemed doctor and scientist, the Neurosurgeon-in-Chief of New York-Presbyterian/Weill Cornell Medical Center and founder and Chairman of the Weill Cornell Medicine Brain and Spine Center.

The description the podcast from December 27, 2019, “Scientific Advances in Brain Reading,” reads: “Artificial intelligence is ushering in a new era of mind reading, with advanced brain scans revealing much of what we’re thinking about. Dr. Marvin Chun, Professor of Psychology and Neuroscience at Yale, explains how researchers in his lab are using fMRI and other new technologies to see what’s going on in your brain, even as you sleep. This holds great promise for those in persistent vegetative states, even as it raises ethical questions about just who gets to read your mind.”

Dr. Chun mentioned a study done in 2011 in which people had their brains scanned using an fMRI machine while they were watching a video. There were several different videos. Using artificial intelligence to analyze the scans, the AI subsequently could look at a novel scan and predict which video was being watched!

This led to a discussion of research in his lab where people are trained to fall asleep in an fMRI machine – and the scans were analyzed to describe their dreams!!!

Another area of discussion was unconscious prejudice. Could we all have our brains scanned, and then have the scans analyzed by AI, to figure out who is prejudiced when picking people for a jury!

What I am curious about is what the scans look like. What data do these neuroscientists look at, and distinguish, to learn about personality, vision and … thoughts? The method of reading the brain scan data is called neural reconstruction. Does it need to be done by AI, or is it based on human determination? Here is a video on the background of this research. It states that there is a “massive database of brain activity” that can be used to “infer what a person is thinking.”

I feel like a detective, trying to find out “hidden” information so that I can figure it out. Will this help me understand how AI could have a positive benefit? Furthermore, just as I am not a computer scientist or engineer, I am not a scientist with a PhD, so I may not be able to comprehend the meaning or impact of the research. I am afraid that it can cause harm if it truly can read our minds. And even sadder, that our spirits can be reduced to finite sine waves.

However, it is still quite fascinating to think that there are some recognizable “thoughts” that can be recorded magnetically. How on earth could this research be used to “prevent” learning disabilities or dementia, as the video states? I wish I could see that connection.

From what I can gather after a few days of looking at articles online is that this particular research is “mining” data and making some connections with both physical and semantic (meaning-making) parts of our brains. AI is needed as a massive processor but not really as an interpreter.

 

Positive aspects of AI

“GOOGLE AI BEATS DOCTORS AT BREAST CANCER DETECTION – sometimes...” https://www.wsj.com/articles/google-ai-beats-doctors-at-breast-cancer-detectionsometimes-11577901600

Business: getting new products, like computer chips, to market faster, and reaching appropriate customers: https://www.wsj.com/articles/intel-hires-hpe-technology-chief-as-new-cio-11578949807?mod=searchresults&page=1&pos=1In its most recent report issued in April, Intel’s IT department used machine learning and analytics to improve the time-to-market for products such as chips by 52 weeks, thereby generating about $2.85 billion in additional revenue, the company said.”

AI for #MeToo: https://www.caltech.edu/about/news/ai-metoo-training-algorithms-spot-online-trolls

Caltech is collaborating with Amazon Web Services to offer scholarships for students to learn how to write AI programs. http://www.eas.caltech.edu/news/1146Pasadena will host part of the AWS Computer Vision Science team, which is creating new algorithms to infer properties of the physical world from images. Such algorithms are fundamental to computer vision and machine-learning applications. These algorithms will be used to improve AWS's current services aimed at detecting and recognizing objects and events in images and video, and will also be used to develop new services that make computer-vision and machine-learning tools available to developers, according to AWS.”

“Explainable” AI (XAI): http://www.cms.caltech.edu/events/87450