Home > data science, modeling > Algorithmic Accountability Reporting: On the Investigation of Black Boxes

Algorithmic Accountability Reporting: On the Investigation of Black Boxes

December 5, 2013

Tonight I’m going to be on a panel over at Columbia’s Journalism School called Algorithmic Accountability Reporting: On the Investigation of Black Boxes. It’s being organized by Nick Diakopoulos, Tow Fellow and previous guest blogger on mathbabe. You can sign up to come here and it will also be livestreamed.

The other panelists are Scott Klein from ProPublica and Clifford Stein from Columbia. I’m super excited to meet them.

Unlike some panel discussions I’ve been on, where the panelists talk about some topic they choose for a few minutes each and then there are questions, this panel will be centered around a draft of a paper coming from the Tow Center at Columbia. First Nick will present the paper and then the panelists will respond to it. Then there will be Q&A.

I wish I could share it with you but it doesn’t seem publicly available yet. Suffice it to say it has many elements in common with Nick’s guest post on raging against the algorithms, and its overall goal is to understand how investigative journalism should handle a world filled with black box algorithms.

Super interesting stuff, and I’m looking forward to tonight, even if it means I’ll miss the New Day New York rally in Foley Square tonight.

Categories: data science, modeling
  1. Brad davis
    December 5, 2013 at 8:24 am

    Your blog makes me wish I lived in New York, along et as much as anything else. Anyhow, I think that the overall question of how to get investigative journalism to work in a world filled with black boxes is similar to my ideal about how science journalists should work if they don’t have enough background in something, and it is the same approach I use. I ask someone who knows more about the topic than I do for help. In the case of journalists, I think it would be good / useful if they would talk with graduate students through postdocs for help to understanding what it is they are reporting on. The student/postdoc would have their work overseen by someone more senior (maybe not for a postdoc) to ensure completeness. This would help scientists with communicating effectively with the public, and would help out investigative journalists (or any, really) build up a network of contacts to go to for help on a topic, and he student / postdoc,could perhaps share the junior by line.


  2. December 5, 2013 at 11:40 am

    I hope to be able to watch it today, have the webcast on my schedule:) It was good yesterday to see the President address inequality too as it’s like the big white elephant in the living room. I just keep kicking along trying to catch as many layman as I can to get the basic message across to address the root of models and algorithms clashing and intersecting in their lives everywhere they turn. It’s what runs on the servers 24/7 that makes life impacting decisions about all of us.

    I keep telling all that the health insurance situation is just like the markets anymore with spasmodic and clashing algorithms searching each other out and executing as programmed, just a a little different focus of course on how risk is manipulated but same bottom line outcome desired, profit. That’s ok if all are honest though and we have efficiencies but as well know that get skewed in some fashion. So in healthcare it’s also the continuous rise and fall (errors & flaws) of the machines, Health insurers sure are rushing to hire both Quants and Data Scientists (I say both as I see all Quants being Data Scientists but all Data Scientists are not Quants) like crazy if you look at classifieds so I’m glad that the effect of algorithms outside of markets is getting attention for sure.

    A couple years ago I had a big Hedge Fund guy spam my blog trying to do one of those algorithmic “reputation restore” programs..it was actually funny and I called him up on it and confronted him as the program created all kinds of “garbage text” with garbage automated blog farms too. Now there’s a case of some real garbage algorithms out there and the junk text was so bad that it made him look like a real fool, it was a good one. He was trying to avoid his name being connected to “Madoff” on Google searches…(grin) and it didn’t work. Oh well one for the archives about algorithms touching us from directions we least expect:)

    I referenced you on my post yesterday as well, so keep it going!


  3. December 6, 2013 at 8:59 am

    We’re seeing the implementation of the apparatus that allows for information manipulation and control at global scale. The ubiquitous, illegal surveillance and data storage in perpetuity by the NSA and their like is only the tip of the iceberg. Eli Pariser exposed the ‘filter bubble’ – but it goes beyond that. it’s what I call the ‘rise of the automated information war’ – http://dharmabuilt.com/the-rise-of-the-automated-information-war/


  4. December 6, 2013 at 12:10 pm

    I had to go out on a job at a doctor’s office but watched last night when I got home, very good! I love your focus on the consumer as we need that. By the way what insurers give the regulators in the states about rate increases is not entirely open either and they pick and choose what they give to insurance commissioners. Some state legislatures are getting smarter and I wrote about that about a year or so ago when New Jersey said to the carriers’ we want to see your algorithms”…in other words not just report with a lot of text to read, so it’s starting.

    Back to a post you made a while back about weight, I’m all over that one too as you have insurers like United and Blue Cross that now buy your Visa and Master Card expenditures from the banks! It was pretty damn weak what Blue Cross said when they were asked why are they buying that data…”we are checking to see if our members are buying clothes a size larger”…please! We all know once they have that data it’s up to their query masters as to what they analyze…and you know what, you could be buying clothes for anyone and so the error factor with that analogy is shot to hell:) Blue Cross had to come up with some kind of an answer is how I saw it, and this is over the edge, so all should be aware of what they are doing and also take notice of all the quants insurance companies are hiring, and who knows what kind of models are being created as they look for profit in any way they can.


    Keep in mind that United as an example has 15 pages of subsidiary companies listed with the SEC and they are in businesses you would least expect, as they are currently firing Medicare D doctors all over the US, but at the same time they sock $50 million into another low housing project in Minnesota, well that could pay a hell of a lot of medical claims from the doctors they are firing who take care of seniors. Guess they figure when we all go broke with healthcare they will be there to provide us a low income housing project where we might be able to afford to live and keep their income rolling in:)

    Propublica does a pretty good job with their analytics too but in my own opinion, they could back off a little on the doctor data with receiving payments as that’s minimal with a level of importance compared to what else is going on out there. I had doctors I know here in southern California look themselves up and gee there were a couple that 4 years ago got $1000 or maybe $2000 for a speech they gave..big deal and and I know there are abusers but for the most part those are the exception and not the rule for what’s going on with doctors today. Supreme court gave them a gift earlier this year to where they can now group to litigate against insurers when settling issues. We can thank United for spending all that money for several years to try to protect keeping that clause in their contracts to where it was not allowed, in other words contract said if you have an issue, it’s you (one doctor) against us, the corporation.


    That case was right in New York and played out with the court in June of this year. Many of the doctors right now have contracts that are paying them at rates less than what Medicare pays and they do this with contracts that have formulas that are complex and difficult to understand, created by lawyers so if doctors don’t sign, they are out of network. That’s my only issue with Propublica, they can downplay that data a little bit as it actually ends up hurting the average doctor with public perception, who’s getting beat up everywhere else they turn.

    Keep up the good work and focus on consumers as we need it and like you I’m tired of seeing technology used to make the wealthy folks richer while using math and algorithms to deny access and take from the poor and middle class, and I’m in there with everyone else!


  1. No trackbacks yet.
Comments are closed.
%d bloggers like this: