Archive

Author Archive

Thoughts on a Hippocratic Oath for data science

I was interviewed by Tom Upchurch of Wired UK on a Hippocratic Oath for data science and ethics of AI (it’s long!):

To work for society, data scientists need a hippocratic oath with teeth

Categories: Uncategorized

I don’t care about self-driving car ethics

My newest Bloomberg View piece just came out:

Don’t Worry About the Ethics of Self-Driving Cars

Road deaths create an incentive to care that is woefully lacking elsewhere.

 

To read all of my Bloomberg View pieces, go here.

Categories: Uncategorized

Teaching Computers to be Fair

My newest Bloomberg View piece just came out:

How to Teach a Computer What ‘Fair’ Means

If we’re going to rely on algorithms, we’ll have to figure it out.

 

To read all of my Bloomberg View pieces, go here.

Categories: Uncategorized

Boring Exercise Update (please don’t read)

Hey all! I told you I’m boring now, so here’s proof. I ran my second 5K race this morning.

My first one was in early February, and I ran with my buddy Elizabeth (who you might remember as my yarn whisperer) and my husband Johan:

B3A88BC4-2A5E-4979-8FC6-E0277BCEDAFF

They gave us hot cocoa and bagels afterwards, which was nice because it was 17 degrees.

This morning my husband watched me run the same route but without cocoa:

E7A8C381-FFC6-46A6-A3A3-A55585A88AFD

This morning it was a balmy 33 degrees at race time.

My goal is always simply to finish the race without stopping, which I’m happy to report I did both times. I run really slow, though, so it always becomes this thing where I’m near the very back, with people who run by me and then stop to walk, and then I slowly pass them, and then they run by me again. By the end of the race I’m kind of the last always-runner and ahead of most of the sometimes-walkers.

I’m happy to report that this second race was easier to complete (I never felt like I wouldn’t be able to, whereas in the first race there were definitely moments I was forcing myself to keep going) AND my time went down! From 39:00 to 38:55, so a full five seconds. For those interested, that’s 0.21% faster. Did I mention that I’m a slow runner? I’m also a very consistently slow runner.

I’ve signed up for a triathlon at Lake Welch in May. It’s a sprint triathlon, so the running part of it is a 5K, which is by far the hardest part for me. So I’m well on my way to training. For the next two months I’ll be doing lots of biking and running and trying to sometimes do two in a day, to build up stamina. The swimming is super easy so I barely need to practice that, but I will anyway because it’s nice sometimes to do something easy instead of forcing yourself to do something hard. Having said that, I’ve really been enjoying the running, it gets out a good portion of my urban aggression. Also I honestly enjoy being that slow runner whom everyone passes, because it makes them feel really fast, and I’m happy for them, and I don’t mind at all being slow. I’m just grateful to be healthy.

OK, sorry to bore everyone with all of this, but if you’re doing similar stuff, feel free to sign up for the same triathlon as I am and I’ll see you there!

Categories: Uncategorized

The Bail System Sucks. Algorithms Might Not Help.

My newest Bloomberg View piece just went up:

Big Data Alone Can’t Fix a Broken Bail System

Philadelphia should think twice about its risk-assessment algorithm.

 

For all of my Bloomberg View pieces, go here.

Categories: Uncategorized

Ernie Davis: The Palantir of New Orleans

Did you hear that Palantir, the big data company founded by Peter Thiel, has been secretly building predictive policing algorithms for New Orleans?

Well, when my buddy Ernie Davis found out, he decided to write a poem. He’s generously allowed me to reproduce it here:

Six years ago, one of the world’s most secretive and powerful tech firms developed a contentious intelligence product in a city that has served as a neoliberal laboratory for everything from charter schools to radical housing reform since Hurricane Katrina. Because the program was never public, important questions about its basic functioning, risk for bias, and overall propriety were never answered.— Palantir has secretly been using New Orleans to test its predictive policing technology, Ali Winston, The Verge, Feb. 27, 2018.

In Eldamar, so long before
Our time, that none can tell the count in years,
The elven craftsman Féanor
Devised the seeing stones, the Palantirs.

The men of old, in seven towers,
Installed the stones that Féanor had wrought
And used their extrasensory powers
To see far off and to converse in thought.

But using a device whose might
Exceeds your wisdom risks a fearful fall.
The fates of Saruman the White
And Steward Denethor are known to all.

********************************************

The enterprising Peter Thiel
Built Paypal and became a billionaire.
A man of business nonpareil
But arrogant as Féanor the Fair.

He scorned the college education
That piles useless knowledge in your head,
And so established a foundation
So youths could start up businesses instead.

He scorned the privileged elite,
Self-righteous, over-educated, smug,
And thus endorsed the loathsome cheat
Who honors every autocratic thug.

Since folks online are always willing
To publish on the web all they can tell
Thiel saw that he could make a killing
By mining it for content he can sell.

His team of workers then designed
The mightiest program they could engineer
To sift the data to be mined.
He named the company “The Palantir”.

The palantirs of Féanor
Could show what was long past and far away.
Thiel’s Palantir sees vastly more:
It knows right now what men will do some day.

It studies billions of relations
‘Twixt men as they develop over time.
And finds the key configurations
That augur the committing of a crime.

********************************************

To prove, past reasonable doubt,
Who’s guilty of specific criminal acts
Requires reasoning about
An awful lot of pesky little facts.

Who was where and when and why?
What show the footprints, blood stains, DNA?
An inconvenient alibi
Can ruin any prosecutor’s day.

A human being is still needed
To comprehend these kinds of evidence
No AI program has succeeded
In mastering the basic common sense.

But building an AI detector
For criminal propensity’s no sweat.
You map a person to a vector
And classify it with a neural net.

********************************************

New Orleans, fair but troubled Queen
Has not in full recovered from the blow
Dealt by Hurricane Katrine
In storm and flood, a dozen years ago.

Gangs that trafficked in the sale
Of heroin and methampetamines
Fought turf wars, and they left a trail
Of murder on the streets of New Orleans

James Carville, famed politico,
Lived in New Orleans and held it dear
And Carville also chanced to know
About the products built at Palantir.

Carville convinced the company
(He was a paid advisor at the time)
To share their software, all for free
To help N.O.P.D. to combat crime.

An altogether secret deal:
Only N.O.P.D. and the mayor,
The folks at Palantir and Thiel
Were any time informed or made aware.

“Fool!” Thus Saruman was named
For secret hid from Council long ago.
And should not those be likewise shamed
Who thought the city council need not know?

Policing with predictive code
The guardians of security delights,
But leads us on a risky road
Toward bias and ignoring civil rights.

Matalin, James Carville’s wife
Assures us all that we will be OK
As long as, in the course of life,
No cousin or acquaintance goes astray.

So Palantir in place remains
And now that we have heard of it, we must
Conclude that those who hold the reins
Of power have betrayed the public trust.

This is part of the collection Verses for the Information Age by Ernest Davis

Categories: Uncategorized

A natural experiment for Facebook

I wrote a Bloomberg View piece about something Facebook could probably do to determine if the Russian ads made a difference in the 2016 election:

Facebook Could Do a Lot More on Trump-Russia

The company could probably measure the effect of election-meddling if it wanted to.

 

For my complete list of Bloomberg View pieces, go here.

Categories: Uncategorized