Home
> Uncategorized > Facebook, the FBI, D&S, and Quartz
Facebook, the FBI, D&S, and Quartz
December 8, 2016
If you’re wondering why I don’t write more blog posts, it’s because I’m writing for other stuff all the time now! But the good news is, once those things are published, I can talk about them on the blog.
- I wrote a piece about the Facebook algorithm versus democracy for Nova. TL;DR: Facebook is winning.
- Susan Landau and I wrote a letter to respond to a bad idea about how the FBI should use machine learning. Both were published on the LawFare blog.
- The kind folks at Data & Society met up, read my book, and wrote a bunch of fascinating responses to it.
- Nikhil Sonnad from Quartz published a nice interview with me yesterday and brought along a photographer.
Categories: Uncategorized
mathbabe 
Where do prefer to have discussions about your pieces? I’m looking forward to what people thing about Facebook and democracy.
LikeLike
The one about the AI bots for the FBI caught my eye right off the bat because that’s what Communist China does with real people. They have thousands of people who troll the internet to find sites/threads on social media that are spreading stuff the CCP doesn’t want spread, so their real human bots are already do what the FBI AI bots would do if they ever came into use and actually did what they were intended to do.
But if AI’s learn and grow through interaction, they could all end up being future conspiracy theory nuts just like Donald Trump and most of the members of his proposed cabinet.
In fact, China’s CCP goes as far as to hire specific types of human bots to deal with different challenges on the Internet that they don’t like For instance, when the CCP declared war on Internet porn sites, they hired thousands of mothers to troll for those sites so the CCP could shut them down. That failed because as soon as they shut one site down, several more would appear to replace it. It was an uphill battle where every step forward ended up taking three steps back.
LikeLiked by 1 person
My mind was blown. Yesterday, I was re-reading The Meaning of it All, by Richard Feynman. Starting on page 89 he lays out the potential for abuse and the fears for society with regard to “Big Data”. He doesn’t use that phrase of course, because he’s giving these talks in 1963. He also uses a lot of words the last evening of the speech to talk about politics, the misuse of science by journalists, what people go through to change their beliefs…among other diamonds.
His book could have been written this year.
And today, I’m catching up on your blog.
LikeLike
Feynman had a very low opinion of social “science” in general — basically what “big data” is all about. In that book he also talks about how social scientists regularly fool themselves with statistics.
As he said, “The first principle is that you must not fool yourself — and you are the easiest person to fool.”
I’d have to agree. Most of the people doing these big data projects (teacher value added is a perfect example) are simply fooling themselves into believing whatever it is they are trying to “prove” with their big data. They cherry pick and ignore and/or reject data that does not fit their model.
What they are doing is most certainly NOT science. I’d have to say that “data science” is actually an oxymoron.
LikeLike
I agree. I’m fond of Ioannidis notion of a “null field” to describe this. Unfortunately, most of the administrative work at my college, for example, is absolutely exploding with this kind of data assessment.
http://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.0020124
LikeLike
Interesting paper.
One thing that Ioannadis talks about is “hypothesis generating” studies. (as opposed to “hypothesis testing” studies)
In the maining of it All, Feynman was actually very critical of so-called “studies” that use the same data to (supposedly) test a hypothesis as was used to formulate it.
As Feynman pointed out, you have to perform ADDITIONAL (randomized, controlled) experiments to actually test a hypothesis to “confirm” (or refute) an apparent “effect”.
It’s actually nonsensical/absurd to use the same data that the hypothesis was based on to to show the “reality” of an effect.
But people nonetheless do it all the time.
LikeLike