Since New Year’s day a slow drip of many documents concerning the work of Cambridge Analytica across 68 countries is giving insights in how the combination of consumer tracking and targeted adverts is being used to influence democratic decisions. Not just within a country, but across multiple countries and simultaneously (meaning foreign interests presented as domestic opinions of the electorate in multiple countries). It’s not entirely surprising, these are age old instruments of propaganda, provocation etc, being redeployed in the digital age, which allows an entirely new level of scale and granularity that makes it a much more malicious beast. It’s shocking on two levels. First, it shows there’s a strong need to make radically transparent to people where material they get served in the silos is coming from, why it is being showed to them, whether it’s part of a/b testing or not, and who is paying/taking influence on each item presented to them. Second, even if there should be no effect at all of these type of campaigns (which seems to crop up as a defence here and there), it is revealing that office-seeking clients and political operatives buy into the cynical premise of the entire concept. Which alone should disqualify them from being elected. The clients need to be held more to account, than the service provider, regardless of any illegality on the side of CA.
The HindSightFiles twitter account is releasing a steady stream of Cambridge Analytica files during the first few months of 2020, leaked by former CA employee Brittany Kaiser. Part of these documents were used earlier in the US Mueller investigation into 2016 election influencing by Russia, and released to the UK Parliament after the initial CA scandal broke.
This Article was mentioned on roel.io
I much like Laura Kalbag’s “I don’t track you” declaration on her blog. She links to that post in the footer of her webpages.
As Laura Kalbag says it’s “as much a fact as a mission statement“. I would definitely like to be able to say the same, because it’s important as a signal, as a statement that the web does not need to be what the silos as advert delivery and manipulation vehicles make it to be. But for this blog it isn’t fully a fact.
I do not track anything anyone does on my site. But others in some instances do. This is the case where I embed material from elsewhere. Although often what I embed is still my own content, such as photos and slides, they are served from the likes of YouTube (Google), Flickr, and Slideshare (LinkedIn). The primary reason for using such services is storage space. Presentations, videos and photo collections tend to be large files, filling up the allocated space in my hosting package quickly. And of course there are occasions where I do want to show content by others (photos and videos). Especially in the case of images, showing other people’s content here is very deliberate, based on an obligation to re-use.
This means that I am an enabler of the tracking that such services do when you visit my blog. To be certain, you have a personal responsibility here too: your browser is your castle, and that Castle Doctrine of browsers means that you should already actively block tracking in your browser. However, I also have a responsibility to not expose visitors to tracking where that can be avoided.
So how to avoid tracking? What alternatives are out there? Here’s a list with the services from which this site over the years has embedded material.
YouTube (Google): I did not know this until I looked for it today, prompted by Laura Kalbag’s blogpost, but Google provides a setting with embedded YT videos that disables tracking and serves the video from a different domain (youtube-nocookies.com). This is what I will do from now on, and I will go through my older postings to change the embed code in the same way.
Flickr: I use Flickr a lot, it’s both my off-site online photo backup, as well as an easy way to post images here, without taking up hosting space. My tracking detection tool (Ghostery) does not find any trackers of embedded images, provided I strip out some of the scripting that comes with an embed by default. This stripping of superfluous stuff I routinely do, and is in my muscle memory.
Slideshare: this I think needs replacing. A Slideshare embed always comes with a Google Analytics tracker and a 3rd party beacon it seems. There is no way I can strip any of that out. It’s a good idea to do without Slideshare anyway, so need to search for an alternative. I might go for my own cloud space, or start making my slides differently, e.g. in HTML5, or find some other tool that I can attach to a private cloud space, and allows easy sharing with others.[UPDATE Oct/Nov 2020: I deleted my Slideshare account and am bringing those presentations to a hosting package and domain I control]
Scribd: this one definitely needs to go too. Embedding a Scribd document adds Google Analytics and a Facebook tracker, and curiously still a Google+ tracker too, though that service no longer exists. Again, need to search for an alternative. Same as with Slideshare. [UPDATE Oct/Nov 2020: I deleted my company’s Scribd account, and am in the process of bringing those documents to a self-hosted environment]
Vimeo: this video embedding service does not add trackers as far as I can tell from my Ghostery tracking monitoring plugin.
23Video: this platform has pivoted to corporate marketing videos and webinars, and no longer supports casual embeds like in the past. I will need to go through my archive though to clean up the postings where I used 23Video.
Qik. This was a live streaming video service I used around 2008. The domain is no longer active, and any embeds no longer work. Will need to clean up some old postings.
So, from this list, Slideshare and Scribd stand out as the ones adding tracking features to this site, and will need to go first. So I’ll focus there on finding replacements. Flickr and Vimeo are ok for now, and Youtube for as long as they respect their own privacy settings. Flickr and Vimeo of course don’t have your data as their business model, whereas YT does, and it shows. Once I’ve removed the tracking functionality from embedded content, what remains is that any call to an outside source results in your IP being logged in that outside server’s logs, and by extension your user agent etc. This is unavoidable as it comes with connecting to any web server. The only way I can avoid such logging is by ensuring I no longer use anything from any outside source, and hosting it myself. For my own content that is possible, as for images I re-use from e.g. Flickr (by serving the image itself from a server I own, and otherwise just linking to the source and creator. As I did with the image below), but hardest for re-using other people’s videos.
Tracks of footprints in the snow, image by Roland Tanglao, license CC BY