The investment will go to multiple projects at the Allen Institute for Technology, or AI2, the Seattle-based nonprofit. In the new “Project Alexandria,” the goal is to create a system imbued with good sense and judgment.
Paul Allen puts a premium on common sense – so much so that he’s investing $125 million to teach it to computers.
The Microsoft co-founder said Wednesday he would commit the money over the next three years to the Seattle-based Allen Institute for Artificial Intelligence, known as AI2.
The funds will go toward multiple AI2 projects, but specifically will be used for the new “Project Alexandria” that will try to bring together various technology elements used in artificial intelligence, with the goal of creating a system imbued with good sense and judgment.
Currently, AI systems can scan and “read” text, interpret some pictures and play board games. But they can’t react to unexpected situations or tell you, say, which way water would flow on a hill.
Project Alexandria aims to teach AI to answer questions such as, “What would you typically find in a trash can?” or “If I put my socks in the drawer, will they still be there tomorrow?”
In some ways, an AI system is smarter than the average child – it can read and store massive amounts of scientific research, for example. But it’s lacking the common sense that most children have, Allen said in a statement.
“If we want AI to approach human abilities and have the broadest possible impact in research, medicine and business, we need to fundamentally advance AI’s common sense abilities,” he said.
This project won’t be something that is perfected and sold within a couple years, said AI2 CEO Oren Etzioni. It takes time and patience to teach machines.
Etzioni compares the problem to having a conversation with Siri or Alexa, Apple and Amazon’s digital assistants. If what you say to them is even a couple words different from the common command, they might not understand.
“They understand your speech, but they don’t understand your meaning,” Etzioni said.
He wants machines to be able to catch simple mistakes – an extra digit added to a patient’s weight at the doctor’s office, for example – that a person would immediately recognize as an error. Otherwise, the machine in this case could order the wrong dosage of medication.
AI2 will start Project Alexandria this year by measuring different AI programs to get a baseline idea of where their common sense stands.
Nonprofit AI2, which has 80 employees, has an ongoing project that teaches computers to answer grade-school science questions and another that aims to make it easier for scientists to quickly do research by sifting through hundreds of thousands of studies in seconds. Techniques learned from studies like this will be folded into developing Project Alexandria.
Allen’s latest investment will about double AI2’s budget, according to The New York Times.