L'Affaire Analytica: Plenty of Snake Oil to Go Around. Category: Editorials from The Berkeley Daily Planet
Publication Date 2018-03-23
Text In the 1980s, when we were toiling in the high-tech vineyards as an unfunded mom-and-pop startup, this quip was making the rounds:
If you look around the table and don’t know who the sucker is, it’s you.
Turns out this maxim originated with poker sharks, but it works for tech too.
In the dustup around Cambridge Analytica, Dr. Alexandr Borisovich Kogan, Facebook and Facebook users, it’s hard to figure out exactly who’s the sucker.
Let’s work our way back to figure it out.
I’d be willing to concede that there’s a fair percentage of suckers in the Facebook user group. As someone who’s been around the software industry since before there was an Internet and even before personal computers, from a technical perspective I’ve never expected that anything I put online was my enduring secret. It’s just too easy to copy code, and after that too hard to control where it goes. Genie back in bottle, etc.
Could Facebook, represented by founder Mark Zuckerberg, be the sucker? Surely not, such a Harvard-admitted Prize Boy must know what he’s doing? If he thought he was protecting the user information he was collecting, should his heartfelt apology be accepted?
Well, when I was a manager in a software company, my motto was that “anyone who believes a programmer deserves what they get.” That quip usually applied to those wild under-estimates by programmers (now elevated to “software engineers”) about when their projects would be finished, but I’m sure it could also apply to guesses by Facebook managers about how secure their data is.
“Not very” is almost always the true answer. Zuckerberg surely knew and still knows that.
And yet… contracts. Those who have access to Facebook data must promise not to use it in the wrong way (whatever that might mean), right?
Yeah sure. My law school Contracts professor hammered on one slogan, perhaps the only thing I remember from that class after some 40 years: “What’s your remedy?” If a licensee has done wrong with your data, what can you do about it after the fact?
Sadly, the usual answer is my favorite Yiddish word, bupkes….nada..nothing. Or as close as makes no difference.
Because money, because time, because a whole bunch of variables can’t be controlled. Lawsuits for breach of contracts like these are very rare, and only happen when the stakes are much higher and the parties are equally matched before they square off.
So then, was there evil intent disguised as naiveté on the part of the academic who had access to Facebook’s user data and sold / gave it away to a for-profit company? Some commentators though so, but maybe he’s really as clueless as he claims to be.
From The Guardian’s profile of the conduit between Facebook and Cambridge Analytica:
“Aleksandr Kogan, a Moldovan-born researcher from Cambridge University, admits harvesting the personal details of 30 million Facebook users via a personality app he developed…While at Cambridge Kogan accepted a position at St Petersburg State University, and also took Russian government grants for research.
“Kogan laughed off suspicions that he is linked to the Kremlin. He said: “This one is pretty funny … anyone who knows me knows I’m a very happy-go-lucky goofy guy, the last one to have any real links to espionage.”
When this story started to heat up, the Internet was awash in speculation about the mysterious professor who had somehow delivered Facebook users to the Trump campaign. Much of this was perhaps fueled by the guy’s propensity to use different names and personas.
According to his Tumblr profile, there’s even a Berkeley connection:
“I have been a psychology professor at the University of Cambridge since 2012 and the founder/CEO of Philometrics.com since 2015. Much of my research focuses on well-being and kindness, but more recently, I’ve been interested in understanding how we can scale survey research without scaling the costs.
“I received my BA in psychology from UC Berkeley in 2008 and a PhD from the University of Hong Kong in 2011. Before going to Cambridge, I worked as a postdoc at the University of Toronto.
“Born in the Soviet Union, I split my childhood half there and half in NYC where my family immigrated. I currently live in the Bay Area.”
No online clue to where he lives now, though quotes in stories seem to reference Cambridge. WhitePages.com reports an address in El Sobrante in December of 2017.
About his name(s). For most of the time, the one he’s used most often, Aleksandr Borisovich Kogan, shows that he’s still got one foot in the former USSR. “Borisovich”, only sometimes included, is his patronymic, not his middle name, and is a version of his father’s given name. His last name, Kogan, is simply a Russian-style transliteration of the common name most English-speakers render as Cohen. Yet on Tumblr he’s called Alex Spectre and at Cambridge he’s been known previously as Dr. Aleksandr Spectre, per his CV as a Psychology Department Research Associate.
The name Spectre seems to have originated with his marriage.
Heavy.com found this explanation on, yes, Facebook: “We chose Spectre as a derivative of Spectrum…We wanted to find a last name tied to light because (a) my wife and I are both scientists and quite religious and light is a strong symbol in both, (b) we got married in the international year of light, (c) we are a multi-ethnic family (so multiple colors of the spectrum) and (d) we just thought it sounded really cool. Changing our last name was a symbol of partnership and unity across racial, cultural and philosophical perspectives.”
But his recent return to “Kogan” might be explained by this item found in California court records, which looks like a sad outcome to a touching tale:
“On 04/10/2017 a Family - Marriage Dissolution/Divorce case was filed by Crystal Ying Spectre against Aleksandr Spectre in the jurisdiction of Contra Costa County Superior Courts…”
If he’d been plain old Alex Cohen all along instead of experimenting with names, he might have avoided a lot of unnecessary drama in media reports of the Cambridge Analytica affair.
It’s not clear amidst all the online debris surrounding his various soubriquets what he actually has been doing for the last few years. A wistful 2017 blog post, The Ph.D.’s dilemma: Academia or Industry?, reveals a bad case of start-up envy which he eventually acted upon.
He wore his academic hat to glean potentially valuable user data from Facebook which he sold to Analytica, which they then peddled to the oligarchic Mercer family to deploy on behalf of a couple of pet candidates, first Ted Cruz and then Donald Trump. The content in these transactions at this point looks an awful lot like junk science, especially as packaged, branded and marketed by Cambridge Analytica.
In the Guardian interview, Kogan himself poo-pooed the value of what he’d sold:
“The accuracy of this data has been extremely exaggerated. In practice my best guess is that we were six times more likely to get everything wrong about a person as we were to get everything right about a person. I personally don’t think micro-targeting is an effective way to use such data sets.”
“It could have only hurt the campaign. What Cambridge Analytica has tried to sell is magic. And it made claims that this is incredibly accurate and it tells you everything there is to tell about you, but the reality is that it’s not that. If you really work through the statistics … those claims quickly fall apart.”
So it was just a lot of cleverly merchandised snake oil?
Well, let’s get back to that Berkeley connection.
In my old email I found a link to a 2011 UC Berkeley press release with this lead:
“There’s definitely something to be said for first impressions. New research from the University of California, Berkeley, suggests it can take just 20 seconds to detect whether a stranger is genetically inclined to being trustworthy, kind or compassionate.”
The lead author of the study was Aleksandr Kogan, a postdoctoral student at the University of Toronto at Mississauga. Co-authors came from his undergraduate alma mater, U.C. Berkeley. Kogan quote: “It’s remarkable that complete strangers could pick up on who’s trustworthy, kind or compassionate in 20 seconds when all they saw was a person sitting in a chair listening to someone talk.”
Indeed it is.
I haven’t checked, but I doubt that this study has been replicated. I’m a believer in Dr. Joyce Friedman’s rule of psychology results: “if it’s not intuitively obvious, it’s wrong.”
It’s a field that attracts snake oil peddlers. Let’s just say that a whole lot of psychology results can’t be replicated for whatever reason . See, further, Wikipedia on the replication crisis in psychology.
If Kogan knew that it was snake oil, and Cambridge Analytica probably did too, who’s the mark at this table? Maybe it was the Mercer family, though Robert Mercer made his fortune doing fancy math for hedge funds and should have been able to figure this one out.
Of course, the pilfered Facebook data could have been mostly wrong but still right enough to tip the scales in 2016. It’s possible that Mercer et al. accurately assessed the small number of voters in the three key states who gave Trump their electoral votes by narrow majorities fueled by tailored “fake news” on Facebook.
It’s a close call, but my candidate for the suckers in 2016 are all those trusting souls who endow commercial social media of all sorts with their intimate personal and political data. I’m reminded of the story of King Midas’s barber, who whispered into some reeds that Apollo had turned the king’s ears into ass’s ears, and the reeds amplified the story far and wide. If you don’t want it spread around, keep it to yourself.
And while we’re on the subject of misplaced trust, the City of Berkeley is now inveigling residents to record their opinions on hot topics with an online app called here, “Berkeley Considers”. Yes, yes, I know it comes with effusive guarantees of anonymity, but it would be oh-so-easy to hack. Voters, citizens, residents should tell their councilmembers and the city staff that they express themselves when they vote by secret ballot in the real elections, period. Quasi-scientific polls like this one are garbage.
Amusing P.S.: when I was googling around with Kogan’s name on a rainy day to figure out what all these clowns were up to, I came upon this example question on a translation website:
“Нужно было вам оставаться в Сан-Франциско, доктор Коган. You should have stayed in San Francisco, Dr. Kogan.”
Good advice—if he’d followed it his professional reputation might have been in better shape today.