In A Twist, Tech Companies Are Outsourcing Computer Work To ... Humans
ARUN RATH, HOST:
Computer programs are super powerful, and they're coming for your job, right? Not so fast. In the past couple of months, Wired staff writer Julia Greenberg has noticed a new hiring trend, starting with a posting for a news editor at Apple.
JULIA GREENBERG: Within the posting, I mean, very specifically Apple says they're looking for humans because algorithms aren't enough.
RATH: Algorithms aren't enough. There are things Apple thinks only a human can do for the company's upcoming newsreader app. And Greenberg says Apple isn't alone. Other giant tech companies are seeking real live humans to do what computers can't.
GREENBERG: I also found similar sentiments on Twitter's website for an editorial lead, which has hinted that it's for their new upcoming event-based feeds. And then we also found some job postings for Snapchat, which for people who use Snapchat, they'll know that there are live stories. And those postings are actually curated by humans.
RATH: And Snapchat has actually poached people from CNN, right?
GREENBERG: Right. So I mean, these companies are very serious about bringing top journalists. And LinkedIn, which has had human curators since 2011, actually poached Dan Roth, who's pretty well-known in journalism communities. He was brought on to be the executive editor to lead LinkedIn's sort of news and content coverage. It's actually journalists who are being sought out for these jobs.
RATH: Not coders, people with actual editorial skills.
GREENBERG: Exactly. Exactly.
RATH: So why do they need the humans? Where are the machines falling short?
GREENBERG: Algorithms are really good at culling through a huge amount of data to bring sort of a more personalized answer, but they're not good at adding emotion. They're not good at showing you sort of, like, what that thing is that you need to know. They're really bad at understanding context in social norms.
And so you'll see with algorithms that there might be two things, like, juxtaposed next to each other that really a human would never put next to each other because it doesn't really make sense, or you might see an algorithm sort of favor certain stories that - for example, this happens a lot on Facebook, where you might click on something out of obligation, but not necessarily because you want to see a lot of that in your feed.
And Dan Roth, who I spoke to at LinkedIn, said that, for example, with breaking news, it's really hard for algorithms to know, like, oh, this just happened. The Supreme Court just announced, you know, this decision. An algorithm might not know that that's valuable. But there would be no doubt that a human who has experience in this would know. And then the other aspect of that is that algorithms are still really bad at understanding humor and language and sarcasm. And so that sort of emotional touch is something that humans can bring to the table.
RATH: So Julia, you and I are journalists and human beings. Is there...
GREENBERG: Yeah (laughter).
RATH: Is there any reason why we just shouldn't think this is completely awesome?
GREENBERG: Well, you know, I talked to some people (laughter) - I talked to some researchers who said to me you know what? Like, right now, algorithms can't do this, but who knows, maybe 10, 15, 20, 50 years down the line. And I mean, it's really important to note that it's not one or the other. The companies are going use a combination of algorithms and humans.
Humans are not going to be able to do that personalized kind of curation, especially at the scale that most of these tech companies operate at. Apple with Apple Music, Spotify and some other, you know, music and entertainment products have definitely also shown that, you know, you need human curation. These are the kinds of thing that you're seeing sort of all across the board.
RATH: As Arnold Schwarzenegger is showing us once again in the movie theater, humans and machines can work together in common cause.
GREENBERG: (Laughter) Yeah, exactly.
RATH: Julia Greenberg is a staff writer at Wired. You can read her piece about humans replacing algorithms that replaced humans at wired.com. Julia, thanks very much.
GREENBERG: Thanks so much. Thanks for having me.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.