Why Elon Musk wants to build a robot maid for you

Elon Musk-backed nonprofit OpenAI announced on Monday that among its technological goals is the creation of a household robot.

Billionaire Elon Musk smiles as he attends a January forum on startups in Hong Kong, China. Mr. Musk his fellow backers of the OpenAI artificial intelligence nonprofit announced plans to develop a housekeeping robot, in a blog post on Tuesday.

Bobby Yip/Reuters

June 21, 2016

Tech guru Elon Musk is known for many things, including his work as chief executive officer of both Tesla Motors and SpaceX, two vastly different companies that are known for their innovation with an eye to the future.

Now, Mr. Musk may be setting himself up for a new reputation – as the man behind squeaky clean floors and perfectly dusted shelves.

Musk’s artificial intelligence nonprofit, OpenAI, wrote in a blog post Tuesday that they are, among other things, "working to enable a physical robot ... to perform basic housework."

Special treatment? How judges are handling Trump ahead of election.

OpenAI was founded in December, 2015, with the goal of creating open-source artificial intelligence to aid humanity. All patents and research generated by the group will be publicly available.

“Artificial intelligence is one of the great opportunities for improving the world today,” OpenAI cofounder Reid Hoffman told The New York Times in December.

Since the nonprofit was created last December, there has been speculation about the projects it will pursue. On Monday, four of the group’s primary backers, including Elon Musk, signed a blog post stating OpenAI’s initial goals.

Among those goals was the modification of a general purpose robot for household chores. Also mentioned in the blog post were the creation of a robotic agent that could perform tasks when asked verbally by a human, and respond with clarifying questions if necessary, and a robot with the ability to solve a variety of games.

Already, many are familiar with Roomba, the room vacuuming robot. OpenAI’s robot, however, would be a more sophisticated modification of an unspecified existing robot for cleaning purposes.

Why Putin’s nuclear saber-rattling on Ukraine sounds different this time

"We believe that learning algorithms can eventually be made reliable enough,” wrote OpenAI, “to create a general-purpose robot."

Pedro Domingos, the author of "The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World" and a professor at the University of Washington, tells The Christian Science Monitor in a phone interview that OpenAI’s quest for a household robot is not unique. 

Many other companies are working on similar projects, he says, but despite longstanding concerns that blue-collar jobs would be the first to be replaced by technology, it is actually very difficult to engineer robots that can handle the multiple tasks required by housekeeping.

Dr. Domingos predicts that OpenAI (and other groups working to create a household robot) will approach the project in steps, connecting pre-existing technology at first, rather than producing a completed household robot right away.

“The things that are easy for us to do,” says Domingos, “are easy because evolution spent millions of years working on them. Things as simple as not bumping into objects on the floor are hard to engineer.”

Monday’s announcement may seem odd, given Elon Musk’s stated wariness of artificial intelligence. In 2014, Musk said in an interview that artificial intelligence is humans’ “greatest existential threat.”

And among the authors of this week’s blog post, Musk is not alone in his concerns.

"I think AI will ...most likely lead to the end of the world,” said Y Combinator president Sam Altman at a tech forum last year. “But in the meantime there will be great companies created with serious machine learning."

Yet, Domingos says that this is highly unlikely, as AI robots don’t have a will of their own and can accomplish only preprogrammed tasks. They can be infinitely smart, he says, and still be harmless.

“The real damage that artificial intelligence can cause occurs when robots do not understand commands. Many robots don’t have common sense,” says Domingos, “and the cure for that is actually making them more intelligent.”

“The real problem is not that machines will become too smart,” Domingos says. “The real problem is that the stupid ones have already taken over the world.”

Regardless of concerns about the potential dangers of artificial intelligence, Domingos says he foresees household robots becoming essential to daily life.

“Having a household robot will become like having a car or a TV,” says Domingos, “I want one. Wouldn’t you?”