Archive for September, 2011

Online Privacy – what you need to know

26 September 2011 Leave a comment

After facebook updated their UI recently, there was a spate of “Do me a favour, hover over my name and uncheck such-and-such a box. I prefer to be private” messages.

Well, let me tell you, you have no hope of retaining your privacy online. Everything you do online is extremely *un*private, and there is very little you can do about it, no matter who checks what.

There are several reasons for this.

  1. Systems change. You might think that you are safe, with the current web site you use. But, then, they change it without your permission, changing features, and, I imagine, prompting the above.
  2. Someone will leak whatever is available to them. Whether it’s deliberately and maliciously, or if their computer is infected with malware, each person/account who can see your posts, photos, whatever, can copy them, save them, and use them later
  3. The website itself can have a security problem. There are countless examples of this, where there is a mistake in a web site, and people can access other people’s accounts, just by a bit of technical jiggery-pokery, or even at random!
  4. The website may have a rogue employee, who harvests some data. I’ve had this happen to me several times, I use a unique email address for most web sites, and a couple of times they’ve been used to send spam to me.
  5. The website can change their terms and conditions at will, and, often, they already own everything you produce on their site – have you checked the T&C for the sites you use? In detail?

This problem has been boiling for years. Eleven years ago I bought and read a book called “Database Nation” about privacy, the electronic trail you leave, and how it would inevitably become impossible NOT to be tracked by databases. Store loyalty cards, automated registration recognition, all store details on a database.

As an example of this, we drove to Portsmouth for an early morning crossing to France, on holiday, a year or two ago. About a month afterwards, I received a letter from a DC, who said that an act of vandalism had occurred the morning we traveled along the M3, and my  car had been one of the last to pass the spot before the incident occurred Рdid I see anything suspicious?

Let’s think about this – the time of passing, and registration numbers of all cars going along the M3 were logged in a database that was available some time after the event. (The letter was not sent for over a month – why sit on the letter, they would have sent it out ASAP, yes?) Given the number of very public security lapses with computer systems, including civilians employed by police forces being dismissed for making irregular and unneeded queries on databases, how safe is this data? And what if this was not a police force, but an online concern that is maximising profit by using unvetted, offshore workers? What if you were suspicious of your other half, and a private eye had a friend who could access data on where cars had been seen? A bit like phone hacking, if you know what to do. Forgetting the police example, how safe is your other info, E.g. email address, credit card, CVC, or, possibly, the password that you use for another 20 sites?

So, almost nothing you do is private. Credit cards, store cards, there are databases everywhere, and data is increasingly being correlated between them. When you start buying organic food, and your store suggests offers on other organic (or other “healthy”) food, this is a single-database action. When you visit lots of guitar-related web sites, and all your social network site adverts start advertising guitar lessons, this is matching across more than one database (well, it’s not, as I’ll explain below, but it serves as an example that people can hopefully understand for now).

Once, my friend went to his local store, and paid (this is back in the days of signature strips on credit cards) with a card the store did not accept, but the teller (poor boy) let him go through the process of swiping and signing. They contacted him (I now wonder how?), explaining they didn’t accept the card, and would he provide details of another one, and he told them to take a hike. Later, he found that they’d later attempted to charge the transaction a different card that he did own, and had used at the store previously. This was for a fairly trivial amount, yet someone was prepared to manually go through credit card receipts, match card names (and, presumably, signatures), then make a false transaction. These days, with everything online, that search would be *so* much easier to do, the match could be done instantly, and on a much higher volume. And it is, not just on store purchases and credit cards. Match your pharmacy payments to your health, so that life insurers get a better idea of how long you’ll live? It’s not far-fetched at all.

So, why is my example of my the guitar adverts above a bad one? It’s all about cookies. Cookies are a way for a web site to “remember” you. The benign view is that it helps you to “remember” the web site, but the balance of power is all on the web site’s. So, youtube, facebook, etc. all remember your login, maybe even auto-log you in, this is all through cookies. There’s a convenience for you, but that’s only part of the story. (Remind me to tell you why free WiFi is bad because your cookies are visible to everyone in the coffee shop when you hit your favourite sites, and people can impersonate you.)

Web sites can be divided up into parts. To the users, they all seem to be part of the same, coherent site, but it can be made up of content from many different sites (often visible on a slow connection, when the page changes shape as different parts arrive). Often the adverts are served from different sites. This is because then the advert-serving sites can count accurately how many views they get, they do not trust the host website to say “Yeah, we served up your adverts 20,000 times today, that’ll be $500 please.” But, although there is no convenience for you, each of those portions of a page can set their own cookies.

Now cookies are a two-way communication. The web site can ask “give me my cookie for the username” and the browser will respond. And the web site can say “give me my secret tracking cookie” and the browser will respond. The value returned, knowing which page you are retrieving, allows the advert site to track you arround. First, you were on the social media site, now you’re home shopping, now a bit of music. The advert site will mke a correlation with your surfing habits, and will serve adverts that are more likely to get notices. So, in this case, it’s not a cross-database correlation, all the correlation is done by the advert-serving site.

And, whenever you visit a site, and there’s a “like” button on it, that link is being made, not just by advert tracking site(s), but by your social networking site.

(As an aside: and, with a new social networking site on the horizon, owned by the biggest advert-serving corporation in the world, what hope do you have if you use that one? )

The bottom line is that you can expect no privacy at all online. It used to be the case that cookies could be deleted in browsers, but there are now “super-cookies” which are much harder to delete. Wherever you go, it’s logged. Your internet address is logged, and this can be used to tie you to a geographical area. Coupled with your browser (web sites can detect plugins, screen resolutions, and all sorts of bizarre stuff), this can be enough to uniquely identify you, especially if you use an unusual browser like me (Opera) with a huge screen (1920×1200 :-)). I’ve given feedback to websites and have them contact me, not with a response, but to ask how I find their website with my particular browser!

And, all your kids are leaving an online presence too. You’re probably doing it for you until they’re twelve or so, but it’s happening. Even in 1998, I used to do an internet search when reviewing job candidate CVs, and in one case found some very (ahem!) material on one candidate. When your kids apply to uni or for a job, their online shadows are going to be searched. They need to be extra-careful what they post. And the internet *never* forgets – web pages are cached by Google, archived by non-profits such as the Wayback Machine, and who knows what criminal organisations can do with the information.

A final thought, it’s getting difficult to recruit police, as every 17-year old has an online presence with images that can be used to identify them. There can be no plain-clothes work for anyone in the days of reverse image search – put an image into a search engine, and it will try to match it with ones it’s seen already.

I can provide references for every assertion I’ve made, but I’m very tired and semi-offline, so digging them out is not easy. If there’s enough response, I will add references, and if you need a particular one, then just post a comment (they are moderated, it won’t appear immediately.)

Categories: online, Uncategorized