When Uber responded to the âGod Viewâ issue by âlimiting accessâ, we had a little chat on Slashdot about whether this was a lie, a good privacy development, or not.
I hypothesized:
So, in a predictable (honestly, surprising they made it to this market cap without doing it already) part of the maturation process; Uber is claiming that theyâll rein in discretionary access to personal information by their frat-bro-asshole management, and instead put full database access to all the data ever in the hands of their advertising and customer analytics weasels.
Thatâs the unpleasant flip side to a story like this. Yes, as it happens, Uber has some of the most punchable management shitweasels one could ask for. The very idea of one of them using âgod viewâ on you makes you want to take a hot shower and scrub yourself until the uncleanness is gone. However, while opportunistic assholerly is repulsive, it is also unsystematic. Once they grow up a bit, and put those data into the hands of solid, value-rational, systematic, people who aim to squeeze every drop of value out of it, then you are really screwed.
Well, Iâd say that those solid, value-rational, systematic, people have identified an area where those data are far more valuable than letting the creeps in management write blog posts about tracking customers who totally got laid, bro!
Iâm not sure if this use is the first(or just one with enough fuzzy spin that itâs relatively public); but Iâm damn sure that it will not be the last, or the most unpleasant.
Youâre absolutely right. And the problem of course is that the more specific the data is, the more valuable it is both to marketing companies, credit companies, and governments both local and national. Yet there is a real potential societal value in using data like this for good instead of evil, with the biggest problem being that there is no transparency to the consumer about how their data is used or who it is sold to for what purpose, or most importantly what the standard is for anonymized data. IMO every tech company should have a clear (human readable) policy about what how they protect their data, how they anonymize it, and who they sell it to. At least then the consumer would be able to decide. And someone out there should be verifying these policies and giving gold stars to companies that have standards in place.
Yup.
My ride to work is out of town this week. Iâve been walking home (itâs under 2 miles) but have been taking a cab into work in the morning because Iâm a lazy ass in the morning (and it was like -2 this morning).
I feel far more reassured when I call the local, established cab company. They have a centralized location, managers I can call if there are issues, and easily identifiable vehicles.
Maybe Uber and Lyft and such will eventually become viable options, but Iâm not comfortable utilizing them in this growing period.
I actually predict that Uber and Lyft wonât be able to maintain themselves as they are now. There will be changes. Eventually there will probably be more hybrid-like services in the future. Cab companies WILL have to evolve, and some already are. I just donât think Uber is going to be a winner, at least in its current incarnation. But time will tell!
ETA: Oh, and the price is about the same. Actually, itâs slightly less than an Uber would likely be, but comes out to about exactly the same with a tip.
What a delicious bit of wordsmithing!
Is the âdataâ bogus and fabricated as the terms âsurge-pricingâ (translation: price-gouging) and âride-sharingâ (translation: for-hire commercial transit) are? And in other news today:
In the second time in two weeks, a driver for the unregulated Uber service has been accused of sexually assaulting a Chicago-based customer:
http://chicago.suntimes.com/news-chicago/7/71/291426/uber-driver-charged-sexually-assaulting-customer
So the cities want the anonymized data for studying traffic flow issues, but it doesnât seem particularly useful for that. Itâs not a representative sampling of all drivers on the road - it is at best a sampling of a particular socio-economic sub-set of drivers, but itâs probably not even that, as people are more likely to use Uber for particular events, not for day-to-day driving purposes. Itâs not going to tell you where drivers are going, itâll tell you what the ride is like for the particular subset of people who have the money to use Uber and a reason to use it. So unless the traffic flow issues they want to study are specifically for the purpose of creating smoother rides just for the wealthier citizens in your city, itâs going to be downright misleading data.
I spotted this on slashdot:
Uber Suspends Australian Transport Inspector Accounts To Block Stings
Accountibility to legitimate government is for the Untermenschen.
Actually it would be pretty good data-youâre not looking for the patterns of uber users themselves, but how long the rides took across certain areas at specific times of day. If crossing boulevard X at hour Y takes 40 minutes when at hour Z it takes 20, that is useful information that applies to everyone who has to cross boulevard X. Your point is valid but a good study can work around it.
Still a gross way of handling user data, though.
While Iâm not terribly surprised by that sort of behavior, just going by past history and all(and the fact that Uberâs management is probably studded with people who think that laws are a conspiracy by the untermenschen to impose slave morality on them), I have to wonder how the valley bros in management got that plan past Legal.
Are âtransport inspectorsâ in Australia legally constituted in some particularly feeble way(such that overtly messing with them couldnât be trivially turned into a few counts of obstruction of justice, maybe with a side of conspiracy, by any halfway displeased prosecutor?) Aside from the bad PR, unless these transport inspectors are magically almost powerless(and canât even call in a police department or something to escalate the matter); this is the sort of âcleverâ plan that could easily turn a bunch of penny-ante fines(quite possibly eaten by your driver peons) into a raft of felony charges aimed at people much higher up the food chain(it may be the driver breaking the law by operating an unlicensed cab; but the little people have no control over user accounts, targeted termination of those is coming from middle management or above).
Also, are they so feeble(and without friends in more dangerous enforcement and surveillance branches) that account banning will stop them for long? If you have to do it all retail, like a peon; buying new burner phones, getting new numbers, new credit cards, etc. is nontrivially expensive and inconvenient. If, on the other hand, you enjoy the cooperation of a telco and a bank(which state agents generally do, so long as the subject isnât âcompetitionâ or âpaying taxesâ) it becomes much more likely that they can provide you with the IMEIs, IMSIs, CC numbers, and email addresses that you need in bulk.
Itâs not as though any of those pitiful little chunks of data are intrinsically expensive, just that obtaining them at retail, punched into plastic cards, adds considerable transaction cost. If you have vendor cooperation you can probably put together hundreds to tens of thousands of âburnerâ IDs programmatically.
Right, but as I said, that data only goes towards improving traffic flow in areas where richer people live and work and the points directly in between. (And not even that, necessarily, as drivers might deliberately avoid driving in lower-income areas.) Because communities in the US are often quite dramatically segregated by income, that can easily leave out large swaths of the city and any traffic âimprovementsâ based on that data can actually make traffic worse for the majority of its citizens. Itâs data with strong built-in economic biases, which makes it bad data.
By âbadâ, you mean âimmensely convenientâ, no?
What could be handier than data that allow you to make Objective Neutral, Data-Driven (totally with science and numbers and these cool computer model printouts the consultants put together!) decisions that coincidentally involve improving things primarily for the people you actually care about; but without the need for, or risk of, any explicitly biased behavior during the process?
I drive a cab and I have driven Uber as well (I use it when the cab I drive is in the shop, or thereâs a special event that will make a few hours profitable but otherwise will suck making it a bad idea to pay for the cab for a full shift, or when I think there will be ridiculous surge pricing). The biggest problem I see from my end with Uber is their misleading rating system. You look at it and see 5 stars and assume that it is a rating of 1-5. The reality is that Uber throws drivers off the system when they go below 4.5 stars. Aside from the fact that it makes the driver rating system misleading (you see a guy with 4.9 or 5 stars and you think he must be awesome, in reality heâs just new because you start with 5 stars and it gets eroded away) since riders are generally unaware of how the system works, the upshot is that even good drivers will inexorably be tossed after a while because:
A) Most people donât bother to rate when theyâre happy, so a disproportionate number of your ratings come from people who are toolbags rating you down for shits and giggles and the occasional unavoidable negative experience (everyone has bad days and makes mistakes from time to time).
B) Even when people who had a good experience rate you they are likely to give you 4 stars, as they believe that this represents a good rating while 5 stars should be reserved for something really special. In reality, by giving a 4 star rating you just moved that driver you liked closer to being kicked off Uberâs system.
The other big issue that arises for me is that the rating system makes it so I have no control over what happens in my car for fear of getting a bad rating. A while back I had a guy get in my car who had just gotten a burrito from a food stand. He asked me if he could eat it in my car, and I said no. He proceeded to eat it anyway. I ended up just letting it go because I didnât want to risk a bad rating and spending 15 minutes cleaning squashed grains of rice out of the seats. Sure, I gave the guy a bad rider rating, but no one looks at those. When the trip comes up on the phone you only have a few seconds to accept it, the rider rating is in a fairly small font not prominently displayed, and youâre usually driving at that moment so you canât take focus away from the road to parse the information.
Convenient if youâre looking to justify decisions made to specifically benefit a particular socio-economic strata, sure. But even then itâs not so useful as commercial and emergency traffic (that benefits those people) and even the routes traveled in the vast majority of day-to-day driving by that socio-economic strata wonât be represented in the data (because youâre most likely to use Uber for specific types of trips). So it may be best for figuring out, for example, how to optimize traffic flow between the homes of richer citizens and their favorite bars (and similar) but not a whole lot else.
I would never get in an Uber, and knowing they plan to sell the data (or will be forced to give the data) to the government, Iâm even less likely to use the service.
Waze data is probably infinitely more accurate and infinitely more valuable, and vast amounts of data about my driving habits belong to google, but I am somehow NOT upset. Iâve thought about it, and I think itâs because I entered into this partnership with Waze because all that data supports me in a direct way. I get the optimum route for me based on my preferences, Iâm made aware of changes in the plan, police, accidents, dead animals in the roadâŚ
yes, until they cross-index it with the personal information theyâre buying from credit agencies and payment processors. then they get path information already indexed by income and social class. there may not be a lot of poor people using Uber, but it happens often enough to get a point estimate out of the data with some averaging.
what, did you think they werenât doing that?
Why the animosity for Uber? I see a lot here. Is it the âstigmaâ of demand based pricing?
Now, as to the article at hand. In my part of the country( and I am sure others ) employees of companies are asked to complete a form detailing how they travel to and from work, distance, the availability of telecommuting, willingness to utilize public transportation, etc.
How is this different? ( with the presumption that customer data will not be shared )
Personally, Iâm far more reassured when I use Lyft or Uber because I know the freaking car is actually going to come. And within a reasonable time window rather than an hour or more. And when it gets there, I know the driver has a GPS that can take me to the address I want to go, rather than asking me for directions! In my experience conventional cab service is broken, at least where I live (San Diego). Iâd use normal cabs if they would provide decent service, but they just donât.