Google's AI Software Beats Humans at Writing AI Software

Not content with jeopardizing other people's jobs, AI researchers are targeting their own.

Mitch Wagner, Executive Editor, Light Reading

January 23, 2017

2 Min Read
Google's AI Software Beats Humans at Writing AI Software

Economists worry that artificial intelligence could make many workers obsolete. Now, AI researchers could be putting their own jobs at risk. Researchers at Google and elsewhere are writing AI software that's better than humans at writing AI software.

Researchers at the Google Brain artificial intelligence research group tasked software with designing a machine-learning system to take a test that benchmarks language-processing software. The software-written software beat previously published results for software designed by humans, according to a report this week on the MIT Technology Review.

And that's not all. Researchers at the nonprofit OpenAI research institute, co-founded by Elon Musk; MIT; the University of California, Berkeley; and Google's other artificial intelligence group, DeepMind, have reported progress in getting learning software to make learning software, the Technology Review says.

One set of experiments from DeepMind could help machine-learning software learn without needing to consume vast amounts of data, the Technology Review says.

But there's a catch: The research requires "extreme computing power," so it's not yet practical for widespread use.

Computer scientists speculate that the advent of AI software that can improve itself could lead to the rapid appearance of a Singularity -- that point where computers achieve greater-than-human intelligence, bringing about the end of the human era. This transition is known as the "hard takeoff" hypothesis, where the Singularity arrives in hours, as opposed to "soft takeoff," where it takes years or decades to come about.

Want to know more about the cloud? Visit Light Reading Enterprise Cloud.

So why haven't the machines replaced us yet?

For one thing, the hard takeoff hypothesis requires "artificial general intelligence" -- machine intelligence that can do anything, as human minds do. Current AI is specialized for particular tasks -- such as language translation or medical diagnosis.

So the machines aren't replacing us today. But I wouldn't make any big weekend plans.

Related posts:

— Mitch Wagner, Follow me on TwitterVisit my LinkedIn profile, Editor, Light Reading Enterprise Cloud

About the Author

Mitch Wagner

Executive Editor, Light Reading

San Diego-based Mitch Wagner is many things. As well as being "our guy" on the West Coast (of the US, not Scotland, or anywhere else with indifferent meteorological conditions), he's a husband (to his wife), dissatisfied Democrat, American (so he could be President some day), nonobservant Jew, and science fiction fan. Not necessarily in that order.

He's also one half of a special duo, along with Minnie, who is the co-habitor of the West Coast Bureau and Light Reading's primary chewer of sticks, though she is not the only one on the team who regularly munches on bark.

Wagner, whose previous positions include Editor-in-Chief at Internet Evolution and Executive Editor at InformationWeek, will be responsible for tracking and reporting on developments in Silicon Valley and other US West Coast hotspots of communications technology innovation.

Beats: Software-defined networking (SDN), network functions virtualization (NFV), IP networking, and colored foods (such as 'green rice').

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like