Monday, March 19, 2018

Facebook's Regulatory Risk Is Real and It Is Magnificent (FB)

So far this year we have had over thirty posts on TheFacebook, as it was once known, with almost all of them being critical of the company in one way or another.
The risks highlighted can be broken down into four broad areas:
1) The media backlash to having their business models undermined by Facebook and Google.
2) The surveillance capabilities of the platform companies and the privacy/security risks they pose.
3) The deliberate addiction of users brains via neurotransmitter manipulation and psychological engineering.
4) The use of the above characteristics by political operators to achieve their own ends and the various outrages, faux and otherwise, elicited.
Our motivation is pretty straightforward: How can we make a buck or two off what appears to be going on?

Recognizing risk before the computers do is one approach. The dirty little secret of machine learning is that the computers can very quickly categorize what they are witnessing only if they have seen the situation previously. We used one of the funnier examples in the outro from December 2017's:

Artificial Intelligence in Risk Management: Looking for Risk in All the Wrong Places
Opportunity is where you find it, turn your risk manager into a profit center. ...
From naked capitalism, November 15:...
...AI cannot cope well with uncertainty because it is not possible to train an AI engine against unknown data. The machine is really good at processing information about things it has seen. It can handle counterfactuals when these arise in systems with clearly stated rules, like with Google’s AlphaGo Zero (Silver et al. 2017). It cannot reason about the future when it involves outcomes it has not seen....MORE 
e.g.
"When Google was training its self-driving car on the streets of Mountain View, California, the car rounded a corner and  encountered a woman in a wheelchair, waving a broom, chasing a duck. The car hadn’t encountered this before so it stopped and waited."
This introduction is getting a bit wordy so we'll chop things into a couple more posts tomorrow but that's the premise, look for risks that AI hasn't yet been trained on in an attempt to gain some asymmetric advantage.

If, in the meantime you can convince yourself you are performing some sort of societal good, all the better. Speaking of the meantime here's some good insight until we get around to parts II and maybe III.
From Motherboard:

Cambridge Analytica's Ad Targeting Is the Reason Facebook Exists
Thousands of third party apps were designed solely to obtain and sell your data. It's no surprise that the data ended up being used again on Facebook, one of the biggest advertising platforms on Earth.

This weekend, while Facebook was quibbling about whether the information used by Cambridge Analytica to target voters in the lead up to the 2016 election was obtained in a data “breach” or somehow using fraudulent means, I decided to check my privacy settings.

Since creating a Facebook in 2006, I have associated my account with 100 separate third party apps. Besides common ones like Spotify, Venmo, and Uber, I have given access to my account to apps like “Typing Maniac,” “I bet I can guess your favorite color,” and “Crazy Cabbie,” among others. Many of these apps I remember only as a faint memory. According to these settings, however, lots of these apps still have access to the same information Cambridge Analytica used to target Facebook users with political ads that helped Donald Trump win the 2016 presidential election.

Typing Maniac—a game I vaguely remember from college—has access to my public profile, my friend list, my relationship status, my “relationship interests,” my birthday, my work history, my status updates, my education history, my events, my hometown, my current city, photos I’m tagged in and that I’ve uploaded, my religious and political views, my videos, my website, my personal description, and my “likes.”...MORE 
Related, January 8's '"Facebook Can’t Be Fixed' (FB)":
Facebook’s fundamental problem is not foreign interference, spam bots, trolls, or fame mongers. It’s the company’s core business model, and abandoning it is not an option.  
More to come.