Jun 14, 2017 08:35 AM EDT
The Most Advanced AI Algorithms Don't Follow Humans At All [VIDEO]
Could Elon Musk and Stephen Hawking be right when they warned the world that AI will soon take over the world and overpower humans? While AI has aided scientists in making the world a better place, recent studies also showed that the most advanced AI algorithms make decisions and perform actions without following a single instruction given to them.
The most advanced AI technology to date is deep learning, a technique where scientists train machines by feeding them different kinds of data. Over time, the machine makes decisions, solves problems, and performs other kinds of tasks on their own based on the data set given to them.
Deep learning is very impressive because it can even diagnose deadly diseases or make million-dollar trading decisions even before humans realize it. One stellar example of this advanced AI algorithm is Deep Patient used by Mount Sinai Hospital in New York.
Data from more than 700,000 individuals were fed to the algorithm including doctors' visits and the patients' test results. When the Mount Sinai team tested its efficiency, the AI algorithm was pretty accurate at predicting diseases based on the patient's records.
But what's puzzling is that it can also predict whether a person is prone to schizophrenia or other psychiatric disorders in the future. The Mount Sinai team wondered how the machine can accurately predict such things since such disorders, especially schizophrenia, is very difficult to predict.
Joel Dudley, who leads the team at Mount Sinai, admitted that they can build such advanced AI models but they have no idea how they work.
The artificial intelligence landscape has dramatically changed as technology advances. The most powerful AI machines no longer depend on commands provided by humans. Instead, they create their own algorithm based on the data and desired output given to them. In short, machines program themselves.
Because of this, there is a pervading fear of who will answer if the machines commit a mistake. How can scientists explain when they themselves do not know how these AI algorithms work?
Join the Conversation