The idea of artificial intelligence knows no limits. But flesh and blood are sweating through the uncertain gig economy behind many of the robots' advances, writes NY Technologist journalist Simon Companello.
"Hi, I'm calling for a customer to book a cutting time, and, huh, I'm looking for a time on the third of May," the human robot says with a weird voice.
Google's demonstration of swallowing an AI service duplex in Tuscany last spring soon became a snack. Have a massage time or book a table at a restaurant restaurant or call a robot assistant who can listen to the call, which has since started in the United States.
But, as we know from numerous examples, today's artificial intelligence is not without its flaws. So when a hairdresser in a loud salon responds to broken English, something comes to Duplex's brain, and someone else needs to run the boat. The savior in distress is, of course, very human. The redundancy in the system is typical of call center workers who take calls that AI assistants cannot handle.
This type of job is called "ghost work", ghost work, anthropologist Mary L. Gray and Siddhartha Suri, who released the book last spring. Ghost Work: How to Stop Silicon Valley from Creating a New Global UnderclassGeneral Chat Chat Lounge It is about people who, in poor working conditions, do all the sweaty work behind the scenes while artificial intelligence stands in the limelight and takes honor.
That's partly. It is about people who are allowed to adjust what artificial intelligence cannot do. Like the Google Duplex substitutes, or Facebook's human-mediated movement that has to decide when robots think decisions are too difficult.
The second category is the people who refine and list all the training data needed to train artificial intelligence. Which is commonly called ot notation.
For example, use the technology behind self-driving cars. It's largely based on image recognition, but computers have no magical understanding of the difference between a bus stop and a cyclist. Or how the slippery road is visually different from the gravel.
So people do a lot of manual work (though companies like Swedish Annotel also want to do this automatically), which is to mark relevant parts of pictures and videos so that computers can learn to see the difference between traffic lights and bicyclists.
Likewise, Google Assistant and Amazon Alexa's language understanding is based on numerous freelance translators and linguists who will teach computers to understand people.
Anthropologists behind Ghost work Is, as the book's subtitle suggests, not an overbowl. He warns that the people behind AI machinery, as part of the so-called gig economy, seem to be into it.
A setting shared by some in the industry. When the Guardian reviewed working conditions internally in a work group for Google's linguists last spring called Pygmalion, a former employee described relationships as slaves.
"Artificial intelligence is not so artificial;" others say in an interview with anonymous translator The Guardian.
Artificial intelligence seems to scare away the fear that we should all be jobless at one time. Instead, we could lose insecure jobs at low wages in the shadow of machines. Ghost's work on an independent basis or with a short contract. Without the benefits of permanent employees, without insurance, vacations, salaries and job security. Instead, we have to wander around the gig economy to find duplicate calls failing to find and save images to ot notet.