In An Era Of Big Data And Total Surveillance, How Do We Protect Our Privacy?
If you haven't heard of Big Data you either haven't been living on this planet for the past few years or you're an ostrich and bury your head in the sand to avoid seeing scary things. Bid data when coupled with Artificial Intelligence can, we are told, tell us how busy the shops are long before the accountants had added up the months takings, how much metal the big warehouses are hoarding or whether office space is in high demand in a particular area.
All of this information can be obtained from just one analytics firm, which gathers information from a multitude of sources, some far more creepy than the almost total sureveillance of our online activity, for example did you know there are fleets of drones, planes and satellites that constantly photograph business premises, car parks, places of entertainment in nations around the world. While those who deal in general information are almost completely invisible you, as an individual, can be tracked via your smartphone either walking around on public transport or in your car if it is a fairly recent model with factory fitted GPS (this is not the reason I drive a 20 year old sports hatchback, a model variant that was designed for production car racing, but lack of modern spyware gadgets is a big bonus.)
A company named RS Metrics sells its customers data based on forensically photographing parking areas and shopping centres facilities from thousands of feet in the air. At regular intervals, its aircraft fly over the target areas taking digital photographs. The company claims the flow of traffic in and out of car parks, business parks or shopping centres enable the generatation of estimates of sales, production or inventories weeks before the official numbers can be calculated.
This may sound far fetched and probably is, but if we think of the blind faith invested in the pseudo - science of mathematical modelling by governments and businesses, such claims probably generate a lot of business.
Taking a more logical and empirical view, to make such clams based on number of cars in the car park or people milling around pedestrian areas seems ridiculous, but businesses are buying it. We can only wonder if the people who buy information from RS Metrics to give them competitive advantage over their rivals ever pause to ask themselves what information RS Metrics might be selling to rivals about their customers.
Everything we look at online, what we buy, our bank transactions, every site we visit and every conversation we have and comment we post is now trackable. Our image can be lifted from private photographs and end up in facial recognition databases, political views we intentionally share only with like minded friends can get us on watch lists, a casual inquiry about some financial service can make us a target for scammers. While we are being submerged under a torrent of information, the personal information than can be inferred from our online activity is being sold to whoever is willing to pay.
The problem for people who like to be well informed is no longer that it is hard to find out something. Instead the problem is parsing morass of data and separating the important from the irrelevant. The uncritical view of Big Tech fanboys is that big data heralds the start of a revolution that will change the way we live and work even more radically than previous Industrial Revolutions have. But human beings are not machines, we are not predictable, therefore the more data you have the more difficult it is to draw meaningful conclusions from that data.
Managing the torrents of available data has become a vital task for maintaining a company's profitability, so much so that it has spawned a new profession, Data Scientists, though nobody seems able to define what a data scientist is. Even with that uncertainly attached to it, data science is said to be driving a new era of economic efficiency. RS Metrics might claim in their sales pitch that developing or buying in the right algorithm can help predict movements the stock and commodity exchanges, companies develop and market the right product and social media sites develop addictive techniques to hold our attention.
Are we prepared for this high tech era in which machines make so many decisions for us? Recent privacy scandals and the public outrage over suicide and self-harm images published by groups hosted on Instagram and Facebook show us blind algorithmic “efficiency” can generate undesirable and sometimes downright dangerous results. Once a website has 'learned' what we 'like', it's software will tweak our profile to give us more of it.
I recall an instance a few years ago when I search using a term on the lines of a Hollywood movie star and American stage actress and the name Tallulah Bankhead who was active from the 1930s to 1950s. For the next few days links to videos billed as " Hollywood star Jennifer Aniston head action," or "see movie star Cameron Diaz give head," proliferated in the results from my searches. I did not check out any of those results, they were obviously fake, but that exemplifies the stupidity of Artificial Intelligence. Not many well ranked results on Tallulah Bankhead but hey, the words Hollywood, star and head appear in his search terms so pages posted by weirdos claiming to show Jennifer or Cameron ‘giving head’ were valid results according to the kind of ‘intelligence’ with which Artificial Intelligence is imbued ...
Some may think the above example is rather coarse. It is, but my search was not, I merely wanted to check a couple of personal details before making an entirely non-sexual reference to Tallhula Bankhead in a fictional work. The point of mentioning it here is to illustrate that AI merely matches locates and identifies character strings in different data sets. And obviously it does not match them very well if “Tallhula Bankhead” + Hollowood is considered to be contextually linked to Jennifer Aniston gives head. Another example of a similar kind of stupidity afflicting Artificial Intelligence happened no to me but to a friend who fell foul of an AI obscenity checker who had an article removed by automatic moderation because of several mentions of a town near to where he lived named Scunthorpe. Scunthorpe can trace the origins of its name back to the medieval era through its being mentioned in the Domesday Book (1086) when the Anglo - Saxon name Skun or Scunn would have referred to a chieftain and thorpe meant a fortified village. Wonderful, marvelous artificial Intelligence could not work that out wheras my humble human intelligence could although I must confess that my surname is Thorpe so I did not have to look up that syllable. But the failure of Artificial Intelligence lies not in its inability to work out the meaning of Scunthorpe but in the fact that it fastened onto a string of letters that in the context have no meaning but taken alone make an obsene term for female genitals. By the same criteria will placed like Blackburn or Blackpool be forced to change their obviously racist names, with Titchfield and Titmore Green be cancelled because the names are sexist and misogyinistic
Instagram, and their parent Facebook might now be trying to tackle the particular issues of self harm, and content encouraging violence while governments are trying to pressure all social media networks and search operators to act to suppress 'hate speech', 'fake news' and 'misinformation' (all of which are terms cannot be legally defined,) but the general trend of using Big Data and Artificial Intelligence to make stupid decisions is, with the wind of media hype behind it, only going to accelerate.
Artificial Intelligence is easily confused, OK it can parse vast amounts of data incredibly quickly and produce lists of matched character strings but is that true intelligence. Throw a metaphor, an ambiguity or anything which requires interpretation at an AI system and it quickly gets its knickers in a twist. Imagine what AI could (or couldn’t make of the phrase ‘gets its kickers in a twist’ unless the programmers had included a list of British colloquialisms in its linguistic references.
One way therefore, to protect our privacy is by reclaiming the art of writing creatively. For years, through education and in the workplace we have been coached and badgered to write and speak in a stilted, colourless way which data scientists deludedly believe will be capable of being mathematically analyzed by algorithms. Fortunately for us data scientists are not aware that context, though not ‘everything’ as a recent UK Prime Minister said, but highly important all the same. Artificial Intelligence can, as mentioned, read millions of words in a few seconds, unfortunately neither the software nor the machine running it has any idea what a single one of those words means. So when we get to inferring the meaning of a word in the context of many others, AI is a non starter.
Artificial Intelligence has been depicted as a monster, a threat to our humanity, but in fact our humanity is the biggest threat to AI. With a few simple precautions we can do a lot towards protecting our privacy. Don’t be predictable, confuse the software as shown above, never be to literal, delete cookies every time you close a browser and set your system to delete browsing history when you shut down the computer AND as you will have heard or read already many times, do not use Google as your search tool. They all track us to some extent but Google’s tracking is intrusive to the point of criminality.
I will not even mention Facebook (Meta), trusting that if you still use any of Zuck’s data piracy systems you have decided to simply not care about the privacy issues.
RELATED POSTS:
Technology rules you
The Conscience of Silicon Valley?
Massive Scandal Brewing In Big Tech