On 4 May 2018 I was giving a keynote on the RTBF Big Data & personalization project in the radio Tarmac studio (see picture below).
The keynote was reserved to RTBF and subsidiaries employees and was well attended. The objective was to raise the awareness on the project I started more than 3 years ago and to introduce topics such as recommender systems, filter bubbles and what public service media can do to improve the contribution of that technology to society.
My story telling was partly based on the recent Cambridge Analytica scandal and the even more recent decision to close the company. I explained that creating more respectful algorithmic tools wasn’t the only solution. We also needed to educate people (and I mean all of them, not only the younger generation) on how data is being used by companies as well as on how it is collected.
As I explained in another article what Cambridge Analytica knew about the millions people it analysed was nothing compared to what data brokers know about us. Opaque companies like Acxiom have decades worth of data on hundreds of millions of people. While some data is extracted from public source (files on bankruptcies, criminal history, …) newest data sources used by Acxiom can prove bery intrusive. Health-related data (not protected in the US) can be retrieved as well as beacon-based location data collected through hundreds of apps that secretely connect to a newtork of beacons throughout the US.
I finished the tour of bad data collection practices by mentioning the infamous Samsung Smart TV case. The audience was stunned to discover that Samsung’s terms and conditions warned of sensitive info being spoken in the vicinity of device and that this information could be shared with third-parties.
I’m more convinced than ever that educating users is the solution. Regulation will help but will remain insufficient as technology and innovation are always ahead of law (read this other article I dedicated to Latanya Sweeney’s keynote at the inaugural FAT conference in New-York this year)Tags: algorithmic governance, recommendation algorithms