Analysis: For artificial intelligence to be 'intelligent,' it needs human trainers

The market for artificial intelligence trainers, or human workers who label and sort data for AI systems, is nine figures and growing, according to a Bloomberg analysis by science writer Matthew Hutson.

For computer systems to perform complex tasks — like driving a car or holding a conversation — human workers must first label large-scale databases of unsorted information. This work is particularly helpful for machine learning, an approach to AI that is taught by experience.

A 22-year-old college student named Katharine Rubin told Mr. Hutson she spends 10 to 30 hours each week evaluating search results and chat responses for 3 cents to 15 cents per task. She's a "part of a growing workforce that spends anywhere from 5 minutes to 40 hours a week increasing the I in AI," Mr. Hutson wrote.

Today, there are more than 1 million people participating in this work, and venture capitalists have poured roughly $50 million in investment funding into startups that connect individuals with AI training tasks.

"For an autonomous car to recognize pedestrians and stop signs, it's typically fed thousands or millions of photos, all hand-labeled," Mr. Hutson wrote. "To nail a conversation, a digital assistant needs to be told over and over when it's failed."

Click here to view the full article.

More articles on artificial intelligence:
STAT: Why are hospitals deploying Watson? To boost reputation
Analysis: Watson for Oncology was over-marketed, now underwhelms hospitals and oncologists
'Robo-doctor' platform leverages AI, blockchain for patient care

© Copyright ASC COMMUNICATIONS 2017. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.


Top 40 Articles from the Past 6 Months