A reader writes:
After reading Marc Ambinder’s summary, I am hoping people aren’t making the same mistake about PRISM that I once made about Gmail. When Gmail first came out, I was working in the California legislature, and a co-worker and I thought it was a terrible idea for Google to, in effect, “read” everyone’s mail and provide ads targeted to them. Our boss introduced a bill to prohibit Google from doing this.
I was assigned to defend the bill at a tech conference, and let’s say I had some misconceptions firmly and uniformly corrected.
No one at Google reads (or could read) anyone’s email. That would be (a) impossible, given the volume of email, and (b) a pretty stupid thing for a company to try to do. Google has pretty sophisticated algorithms that can scan millions of texts for words and phrases that advertisers believe would be relevant to a particular commercial purpose. Ads matching those terms are posted next to the email, and no human (except the recipient) has ever seen anything.
I’m not sure if any actual humans ever see any Facebook postings, but my guess is that the first pass of PRISM works like Gmail. Someone has developed algorithms for potentially dangerous words and phrases, and the millions or billions of Facebook posts are scanned for those. The algorithm’s bar would have to be fairly high, since the number of posts would be astronomical, I would imagine.
Posts that make it over the bar (still not having been viewed by any human being) would then be collected into some output that IS more closely examined, and this may be the stage where humans might be involved. Again, I don’t have any special knowledge here, but I honestly can’t imagine how this could work any other way. The only things that are ever actually seen by human eyes are those that have some markers of potential serious threats.
I can see how some people might still find little comfort in that, and I’m sure there would have to be many false positives in a system like this. But I think it’s far more consistent with your intuition about why this isn’t such a horrible invasion of privacy – an intuition that it seems a lot of us share.
That difference between technological review of data and human eyes viewing (and possibly abusing) communication is an important distinction. If PRISM is more like Gmail than like J. Edgar Hoover’s private FBI files, then this has less to do with privacy than some people might fear. I, for one, got over my concerns about Gmail, and happily got one of its first accounts, which I use to this day.