SubscribeArrow

Have your friends signed up?

What would you like to read about? Hit reply to this email or message me at steve@axios.com. Kaveh Waddell is at kaveh@axios.com and Erica Pandey at erica@axios.com.

D.C. readers: Join Axios' Mike Allen and Evan Ryan tomorrow at 8:30 am for our next News Shapers event.

  • They'll sit down with Oregon Gov. Kate Brown and New Hampshire Gov. Chris Sununu, along with actress Jennifer Garner and Save the Children's Mark Shriver, to talk about how to improve and expand education in America.
  • RSVP here

Okay, let's start with ...

1 big thing: Where the jobs aren't

Illustration: Sarah Grillo/Axios

A swath of millions of Americans have been jobless for a year or more — the hard-core unemployed. Among the causes for their stubborn joblessness are lack of skills, drug habits and felony records.

But there is another, largely overlooked reason: Many of these unemployed people simply can't — or won't — get where the jobs are.

Erica writes: Increasing numbers of experts say the concentration of wealth in big cities, along with the rise of automation, is putting low-wage jobs out of the physical reach of workers.

  • Often they live far on the periphery of cities, without easy access to public transportation.
  • And for the first time in the nation's history, big numbers of Americans have stopped moving for work. Even when there are jobs in another city or state, they have been unwilling, for reasons no one has been able to decisively explain, to pick up and start a new life.

Both workers and employers are getting hit.

As we've reported, about 13 million working-age people in the U.S. are unemployed, involuntarily working part time or have wholly given up on the job hunt. The problem, though, isn't a lack of jobs — employers complain that they can't fill open positions — but a slew of barriers.

A new report from the Urban Institute analyzed 2017 application data across several types of low-wage, hourly jobs in 16 major U.S. cities, looking at the mismatch between openings and job seekers by zip code.

  • In 41% of Boston-area zip codes, there were far more openings than job seekers who lived a reasonable distance away. In other words, most possible applicants lived too far from the workplace, and there either was no efficient public transportation to the job or the cost was too high to get there.
  • In New York, 32% of zip codes had this problem. In San Francisco, 29%.
  • The trend stretched to middle-sized cities, too: 24% of Denver-area zip codes had this mismatch. In Columbus, 37%.
"These are people who are looking to enter the job market, and if they can't find it feasible from a transportation or cost or distance perspective, that's saying that there are significant barriers in place."
— Joseph Kane, Brookings Institution

What's happening: Poorer city residents live in the "last-subway-stop" parts of cities, a couple of hours, or farther, from work — and they either can't afford to or don't want to move.

Another bucket of low-wage jobs, those in factories or warehouses, have the opposite problem. Companies like Amazon are building their warehouses far away from city centers, close to airports and highways and away from major public transit lines. They can be difficult to access without a car, says Brookings' Kane.

  • In Minneapolis, Amazon and General Mills have started offering shuttles to ferry workers from the city to the warehouses, reports the Star Tribune.

The solution is twofold, Christina Stacy, lead author of the Urban Institute report, tells Axios. Cities need to revisit train and bus lines and build affordable housing closer to downtown. "Until we figure out how to make every job a high-quality job, connecting people to a job is crucial," says Stacy.

2. The problem with AI medicine

A doctor points out a fracture to a patient, Kemmling, Colorado, 1948. Photo: W. Eugene Smith/LIFE/Getty

One thing artificial intelligence excels at is making predictions from patterns in huge troves of data. In medicine, it's a big strength, potentially allowing doctors to punch through and recommend much better medicine or procedures to treat their patients.

  • Because of the enormous promise, the FDA has fast-tracked devices that use predictive AI.
  • But now, some scientists — wary of rushing too quickly into a new era of computer-aided health care — are calling for more stringent rules before new AI algorithms can be used on patients.

Axios' Eileen Drage O'Reilly reports: Several researchers published a paper today in the influential journal Science calling for new standards for safety and effectiveness before AI algorithms are approved for use on humans.

  • Predictive AI for medicine has come out of nowhere in the past five years, moving at a "staggering" pace, says Ravi B. Parikh, co-author of the paper and a fellow at University of Pennsylvania's School of Medicine. He tells Axios:
"If these tools are going to be used to determine patient care ... they should meet standards of clinical benefit, just as the majority of our drugs and diagnostic tests do."
  • That's because they're making crucial decisions about patients' health. Eric Topol, director of Scripps Research Translational Institute, tells Axios that predictive algorithms can actually harm patients if they are poorly implemented.

Go deeper: Read Eileen's full story

3. Mailbox

Photo: Getty

On our story about Fortune 500 companies owing no taxes:

"I am concerned that the article may suggest to certain individual taxpayers that they have been treated less than fairly by the Trump tax cut. For example, the refund claimed by General Motors is the result of net operating loss (NOL) carryovers from previous years, and not so much new (and undeserved) tax ‘breaks’ from the tax cut.
The NOL deduction intends that a taxpayer with fluctuating income over a period of years not pay more federal income tax than a taxpayer with a relatively constant level of income. For GM, and many other companies, the tax cut reduced the amount of claimed refunds."
— J. Leon Peace, Jr., Silver Spring, Maryland

On our story about keeping AI away from the bad guys:

"Really great stuff on AI.  But as we suppress dual-use AI technology, the Chinese will of course put it to the worst possible use as fast as possible, both internally and for globally disruptive purposes. I think that’s the real dilemma."
— Ed Bergan, Davidson, North Carolina
4. Worthy of your time
Expand chart
Data: Brookings; Chart: Harry Stevens/Axios

The made in the USA comeback (Rana Foroohar — FT)

How to crack down on surprise medical bills (Caitlin Owens — Axios)

A philosopher argues AI can't make art (Sean Dorrance Kelly — MIT Tech Review)

The life story of your supermarket chicken (Robyn Metcalfe — WSJ)

When kids find their whole lives online (Taylor Lorenz — The Atlantic)

5. 1 🤔 thing: Deciphering emojis in the courtroom

Back when emojis weren't a problem. Photo: Howard Sochurek/LIFE/Getty

In a California courtroom last week, prosecutors were trying to prove that a man caught in a prostitution sting was guilty of pimping. Part of their evidence was a string of emojis, reports The Verge.

Erica writes: The prosecution put an Instagram direct message between the defendant and a woman in front of the judge. The message read, "Teamwork makes the dream work 👠💰," which lawyers said implied a transactional relationship. But the defendant argued the relationship was romantic.

The big picture: It's difficult to decipher what exactly emojis mean — and they rarely sway a jury, per The Verge — but these symbols are becoming increasingly common in court cases. Eric Goldman, a Santa Clara University law professor, told The Verge that just over 30% of all court cases in 2018 had some reference to emojis or emoticons.