Teaching machines not to cheat is the latest project of Google DeepMind and OpenAI

Maybe it's all those Terminator movies, sci-fi novels, and post-apocalyptic video games, but it just makes sense to be wary of super-intelligent artificial intelligence, right?

For the good of, well, everyone, two major firms in the field of AI have convened to work on ways to keep machines from tackling problems improperly — or unpredictably.

OpenAI, a group founded by techno-entrepreneur Elon Musk and Google DeepMind, the team behind the new-reigning Go champ, AlphaGo, are joining forces to find ways to ensure AI solves problems to human-desirable standards. 

While it's still faster sometimes to let an AI solve problems on its own, the partnership found that humans need to step in to add constraints upon constraints to train the AI to handle a task as expected.

Less cheating, fewer Skynets

In a paper published by DeepMind and OpenAI staff, the two firms found that human intervention is critical to informing AI when a job is performed both optimally and correctly — that is to say, not to cheat or cut corners to get the quickest results.

For example, telling a robot to scramble an egg could result in it just slamming an egg onto a skillet and calling it a job (technically) well done. Additional rewards have to be added to make sure the egg is cooked evenly, seasoned properly, free of shell shards, not burnt, and so forth.

As you can guess, setting proper reward functions for AI can take a major amount of work. However, the researchers consider the effort an ethical obligation going forward, especially as AIs become more powerful and capable of greater responsibility.

There's still a lot of work left to go but if nothing else, you can take solace that AI's top researchers are working to make sure robots don't go rogue — or at least, not ruin breakfast.

Via Wired

TOPICS
Parker Wilhelm
Parker Wilhelm is a freelance writer for TechRadar. He likes to tinker in Photoshop and talk people's ears off about Persona 4.
Latest in Tech
Ray-Ban smart glasses with the Cpperni logo, an LED array, and a MacBook Air with M4 next to ecah other.
ICYMI: the week's 7 biggest tech stories from Twitter's massive outage to iRobot's impressive new Roombas
A triptych image featuring the Sennheiser HD 505, Apple iPad Air 11-inch (2025), and Apple MacBook Air 15-inch (M4).
5 unmissable tech reviews of the week: why the MacBook Air (M4) should be your next laptop and the best sounding OLED TV ever
Apple iPhone 16e
Which affordable phone wins the mid-range race: the iPhone 16e, Nothing 3a, or Samsung Galaxy A56? Our latest podcast tells all
The Apple MacBook Air next to the Dyson Supersonic R and new AMD GPU
ICYMI: the week's 7 biggest tech stories from the best tech at MWC to Apple's new iPads and MacBooks
A triptych image featuring the Bose Solo Soundbar 2, Nothing Phone 3a Pro and the Panasonic Lumix S1R II.
5 trailblazing tech reviews of the week: Nothing's stylish, affordable flagship and why you should buy AMD's new graphics card over Nvidia's
The best tech of MWC 2025 examples, including the Nothing Phone 3a Pro, the Nubia Flip 2, and the Lenovo Solar PC
Best of MWC 2025: the 10 top tech launches we tried on the show floor
Latest in News
A man holds a smartphone iPhone screen showing various social media apps including YouTube, TikTok, Facebook, Threads, Instagram and X
Ofcom cracks down on UK tech firms, will issue sanctions for illegal content
Google Chromecast 2
Google rolls out another Chromecast bug fix for users who factory-reset their devices
A Starfew Valley theme on Wear OS
Someone made a Stardew Valley theme for Wear OS and it's perfect
3d rendering of a submarine power cable on the seabed
Subsea internet cables can now ‘listen’ for sabotage using irregular pulses of light
Google Pixel 9 front and back
The Google Pixel 9a has gone up for sale and it’s not even out yet
Panos Panay and Alexa Plus
Amazon's Panos Panay teases future Alexa+ devices from speakers to possible wearables