Online Privacy: make some noise
For personal interest, I am currently researching the topic on biased data and how it affects Machine Learning when data is not cleaned properly. A recent publication showed that Amazon’s recruiting AI preferred male candidates because the machine learned from data collected on hired candidates for tech jobs in the company in the past 10 years (Fun Fact: this application was developed in Amazon’s Development Center in Edinburgh). The argument that this kind of hiring process is inhuman might as this point be invalid anyway. Many big companies have already installed an unsurmountable process of automated evaluation of candidates, especially for graduates, students, internships and graduate schemes. Personally, I avoid applying for companies with such hiring processes simply because I deeply believe, this is not how a team should be formed and I would not be comfortable with the company culture.
So I was wonder how one could create dirty data, add “noise” to our data. My first guide would be Obfuscation – A User’s Guide For Privacy and Protest by Brunton and Nissenbaum (Section 2.14). It explains why feeding Facebook with more data than less is beneficial. Anyway, Facebook will collect data from you, in order to not have less but accurate data, one might prefer to add some ‘noise’ to their personal data. However, it seemed like a lot of work to me, to come up with all kinds of life events and details about a fake life. But for those who don’t want to be profiled and sold out by Facebook, should write their story.
Some simpler tools to use in order to add noise to your browsing history are available. One tool actively opens browser tabs and searches for random things. On the flip side, this will have an effect on the performance of your laptop. Therefore, the inventor of makesomenoise suggests running the application while you sleep. There are other tools, who solve this problem and use python scripts instead to create http/dns traffic in the background. No matter which tool you use to add noise to your traffic data, a machine will not be able to distinguish between real data created by individuals and noise added by the tools. Humans will, however, be able to tell the difference, because the tools generate data that is too random.
On the topic of online privacy, I feel obliged as an Austrian citizen to give an update on activities by Max Schrems. Last year Schrems founded an organisation called nyob, which is short for “none of your business”. The non-profit organisation sees itself as the European Center for Digital Rights. The day the GDPR came into force he filled four complaints. All four complaints deal with the issue of forced consent of big US-tech companies, including Google (Android), Instagram, Whatsapp, and Facebook. Complaints were filled with the CNIL, DPA (Belgium), HmbBfDI (Germany) and the DSB (Austria). In particular, the complaints focus on bundling services provided with the requirement of consent, the necessity of data and annoying pop-ups.
