Skip to main content

You make this possible. Support our independent, nonprofit newsroom today.

Give Now

Paul Allen never created Data, from Star Trek. But he wanted to

caption: Data, the sentient robot from Star Trek, The Next Generation
Enlarge Icon
Data, the sentient robot from Star Trek, The Next Generation

It was February of 2014, and hundreds of thousands of Seahawks fans flooded the streets of downtown Seattle to celebrate their team’s superbowl victory.

But Paul Allen, the team's owner, had other things on his mind: He was sending emails to Oren Etzioni, an artificial intelligence researcher.

Etzioni remembers receiving those emails.

“I thought wow, this is the last time I would think Paul would be thinking about A.I.," he said. "But he was already onto the next thing on his mind.”

The next thing on Allen's mind was the Allen Institute for Artificial Intelligence, which Etzioni now runs.

That spring, in an interview, someone asked Paul Allen how close researchers were to creating an intelligent robot, like Data, from the show Star Trek.

"We're so far away from that level of understanding," said Allen. "But these are the big questions... How could you... create some kind of entity that's actually conscious? Those are the fascinating questions! We all love to be working on those questions."

Allen has since invested hundreds of millions in A.I. research, Etzioni said. That's enough money to fund the 90-person AI2 Institute for years into the future, he said.

"I used to joke that the worst thing that Pete Carroll ever did for me was to take his scrappy team and win the Superbowl," said Etzioni. "Because now Paul Allen expects me to win the A.I. Superbowl."

Etzioni said Allen was speaking with his tongue in his cheek — his way of teasing. "But he also wanted to win," said Etzioni.

That meant not getting distracted by the latest scientific fads. At other academic institutions dependent on fickle funding sources, those fads can sometimes drive the direction of research.

"As you’re going after these long term scientific goals, there are always these sirens," said Etzioni. "Not in the sense of ambulance sirens, but in the sense of Greek mythology. 'What about me?' and 'Wouldn’t this be cool?' and 'Why don’t you work on this short term project?' And Paul was absolutely 100 percent focused on the long term," he said.

Allen's long term goal was to create an A.I. that could pick up a textbook, read through a chapter, and then successfully answer the quiz questions at the end of the chapter.

Data, from Star Trek, could probably do that. And if the place he first did it was at the Allen Institute for Artificial Intelligence, that would have meant Allen had won the A.I. Superbowl.

Paul Allen died on Monday at age 65 from complications with cancer.

Why you can trust KUOW