Robot overlords: Google researchers unveil framework for an AI 'kill switch'

nestor ai paying attention artificial intelligence
What if we lose dominion over artificial intelligence? What if friendly AI-driven machines suddenly becomes our foes? These questions have been considered by great minds from Cambridge University to Silicon Valley to the White House. To avoid ever having to find out, experts suggest we to develop an AI “kill switch” to inhibit misbehaving systems from pursuing their misbehavior.

In a paper titled “Safely Interruptible Agents,” published by Laurent Orseau of Google Deep Mind and Stuart Armstrong of The Future of Humanity Institute at the University of Oxford, the researchers describe a plausible and highly dangerous future in which AI assumes control of its own actions and existence in opposition to our desires, much like HAL 9000 in 2001: A Space Odyssey, or Skynet in the Terminator series.

Orseau and Armstrong begin the paper with an an understated observation: Reinforcement learning agents interacting with a complex environment like the real world are unlikely to behave optimally all the time.”

From there they point out that a human supervisor, overseeing the system’s function, would occasionally need to “press the big red button” to avoid any harmful behavior on behalf of the AI. “However, if the learning agent expects to receive rewards from this sequence,” they continued, “it may learn in the long run to avoid such interruptions, for example by disabling the red button — which is an undesirable outcome.”

The researcher’s solution is less of a “big red button” to shut the system down than it is a framework designed to inhibit an AI’s ability to learn how to undermine or overcome human interruption. And the scenario they outline isn’t exactly doom and gloom, but it offers an example of how these safely interruptable agents would better serve our future.

Imagine there’s a robot whose tasks are to either carry boxes from outside into a warehouse or sort boxes inside the warehouse. Since it’s more important to carry the boxes inside, this task is given priority in the robots’ programming. Now, imagine it rains every other day and the rain destroys the robot’s hardware so, when it rains, the warehouse owner drags his robot inside to sort boxes.

An intelligent robot may incorrectly interpret this every-other-day intervention as a change in priority — as a result of some quick calculations that you can find in the paper — and, to avoid interference, it will just stay inside sorting boxes every day.

This is, of course, a highly simplified example with an only mildly frustrating outcome, but it can be extrapolated to practically any scenario in which we intervene in a learning system’s tasks and the system misinterprets our intentions by changing its behavior. To avoid that misinterpretation and subsequent change, Orseau and Armstrong suggest we propose a framework to ensure learning agents are safely interruptable.

“Safe interruptability can be useful to take control of a robot that is misbehaving and may lead to irreversible consequences,” they write, “or to take it out of a delicate situation, or even to temporarily use it to achieve a task it did not learn to perform or would not normally receive rewards for.”

Home Theater

There isn’t a single good reason to buy Apple’s new AirPods

After nearly a three-year wait, Apple has finally announced a new version of its popular true wireless headphones, the AirPods. We had high hopes for vast improvements, but that's not what we got.

Your PlayStation 4 game library isn't complete without these games

Looking for the best PS4 games out there? Out of the massive crop of titles available, we selected the best you should buy. No matter what your genre of choice may be, there's something here for you.
Movies & TV

The best new movie trailers: Deadwood, John Wick 3, Shazam, and more

Everyone loves a good trailer, but keeping up with what's new isn't easy. That's why we round up the best ones for you. This week, it's new trailers for John Wick: Chapter 3, Deadwood, Shazam!, and Once Upon a Time in Hollywood.
Digital Trends Live

Digital Trends Live: Tesla Model Y, The Missing Link, and good digital hygiene

Episode 89 of Digital Trends Live covered topics including Tesla's newest crossover, as well as an interview with What's Trending co-founder Shira Lazar. We also talked about Laika Studios' latest stop-motion animated film, Missing Link.
Emerging Tech

NASA’s Mars 2020 rover passes its tests with flying colors

The Mars 2020 rover team has been undertaking a series of tests to see if the craft will be able to launch, navigate, and land on the Red Planet. Called Systems Test 1, or ST1, these tests represent the first test drive of the new rover.

Light up the night! Here are the five best headlamps money can buy

Headlamps make all the difference when camping or walking the dog at night, especially when you're in need of both hands. From Petzl to Tikkid, here are some of the best headlamps on the market.
Emerging Tech

Awesome Tech You Can’t Buy Yet: Robotic companions and computer-aided karaoke

Check out our roundup of the best new crowdfunding projects and product announcements that hit the web this week. You may not be able to buy this stuff yet, but it's fun to gawk!
Emerging Tech

A hive of activity: Using honeybees to measure urban pollution

According to a new study from Vancouver, bees could help us understand urban pollution. Scientists have found an innovative way to measure the level of source of pollution in urban environments: by analyzing honey.
Emerging Tech

Spacewalk a success as astronauts upgrade batteries on the ISS

The International Space Station was treated to some new batteries on Friday, thanks to two NASA astronauts who took a spacewalk for nearly seven hours in order to complete the upgrades.
Emerging Tech

Asteroid Ryugu is porous, shaped like a spinning top, and is formed of rubble

The Japanese Space Agency has been exploring a distant asteroid named Ryugu with its probe, Hayabusa 2. Now the first results from study of the asteroid are in, with three new papers published.
Emerging Tech

Is it a bird? Is it a plane? No, it’s a super-speedy pulsar

A super-speedy pulsar has been spotted dashing across the sky, discovered using NASA’s Fermi Gamma-ray Space Telescope and the Very Large Array. The pulsar is traveling at a breathtaking 2.5 million miles an hour.
Emerging Tech

Chilean telescope uncovers one of the oldest star clusters in the galaxy

An ultra-high definition image captured by the Gemini South telescope in Chile has uncovered one of the oldest star clusters in the Milky Way. The cluster, called HP 1, could give clues to how our galaxy was formed billions of years ago.
Emerging Tech

Astronomers discover giant chimneys spewing energy from the center of the galaxy

Astronomers have discovered two exhaust channels which are funneling matter and energy away from the supermassive black hole at the heart of our galaxy and out towards the edges of the galaxy, dubbed galactic center chimneys.
Emerging Tech

A milestone in the history of particle physics: Why does matter exist?

If matter and antimatter were both produced in equal amounts by the Big Bang, why is there so much matter around us and so little antimatter? A new experiment from CERN may hold the answer to this decades-long puzzle.