Algorithms exist all around us. They put our lives on pause and we perform them every day without even knowing it. While they are often judged negatively (that’s because they sometimes affect our personal lives too much), they still exist and can guide and help many practices…
Through this article, you will undoubtedly understand why their use is so controversial, but also why not everything is necessarily black or white (the structured data of the algorithm makes it possible to know the traffic, temperature, especially on motorways. eg. field or very useful possibilities in our daily life).
We’ll also tell you some keys to how they work on social networks as well as some ways to combat them to make them fail you better.
An algorithm is a set of instructions, a series of operations that make it possible to answer a problem or obtain a result.
Let’s take the example of a recipe. To reach the end of our solution and get the desired result, we need to follow a number of steps. Well, the algorithm works like a recipe: it is made up of instructions that must be followed one after the other to produce something or solve a problem. This is a pattern that will repeat itself in a loop.
Algorithms are formed according to a number of parameters such as our personal tastes or our consumption habits. They will try to figure out the repetition of our behaviors, the way we act, our choices, and our preferences in order to prepare themselves.
Algorithms are based on both usage (if you use Waze regularly, the application should recommend your preferred route, for example) and on personal data (you know, the web tracers you use on your internet). Surfing sessions). So an existential question arises: does the algorithm have a link with cookies (the well-known tracers we should now accept on websites) and therefore with the GDPR (in Europe) or the CCPA (in California)? You should get the answer by reading this entire article.
As Aurelie Jean explains in her book Algorithms make lawsThere are two types of algorithms: implicit and explicit.
An algorithm is “explicitly” defined by a set of logical rules to answer a problem.
In contrast, a “built-in” algorithm is a tool whose conditions and rules are generated automatically by statistical learning (we then speak of machine learning).
That’s why we find algorithms in many areas of our lives, from research on Google Maps to Netflix and even Google Ads using machine learning (if that’s one of your acquisition lever techniques) or the use of our own social networks. is for.
Understand social media algorithms
Last February, we conducted a live session about social network algorithms and how to play with them when you need visibility into digital communications.
Algorithms on social networks are organized (implicitly) with the aim of seducing us and spending as much time on the network as possible. For this, as stated earlier, algorithms will create a database to collect all our personal information, consumption habits, choices we make, whether we like to see or not.
After collecting all this information, the algorithm will aim to find the most suitable content for the target audience, so that the user has a good time on the network so that he wants to spend as much time as possible on it. ,
Algorithms on social networks use algorithmic bubbles that represent filtering, with the aim of improving the content that will be delivered according to his preferences and interests to the user who will already be registered (this filtering is mainly done on Facebook). Filter bubbles affect the habits we have adopted in relation to our searches on the Internet.
It’s all the fault of Artificial Intelligence!
The goal of social networks is to have us on their platform, and for that they should do everything to make us happy. This is when these algorithmic bubbles come into play, which manifests on our psychological flaws by offering us content that attracts us and distances us from reality that contradicts our opinions.
We then know that a social network like Twitter will “classify individuals according to their behavior on the social network … such as the type of publications they share, or the content of their own publications” (Aureli Jean).
On YouTube, it is the viewing time (i.e. the video viewing time) that prevails, whereas on applications such as TikTok, it will be the repetition of a short video.
To learn more about the use of these algorithms on social networks, you can consult this debate between Florian Vincent and me.
Last November, two obit journalists exchanged their TikTok accounts to counter the Tik Tok algorithm, which, I quote, “knows me better than my mother”.
Result ?
Overview of an algorithm that “tracks you … that is, every time you pause on a video, that you hesitate, or that you watch it again, the algorithm tracks you”.
Limitations and Dangers of Algorithms
An algorithm is simply a set of instructions, created by humans, and therefore can have flaws. An error can lead to erroneous results, and often these errors are not identified in time.
Recently, many scandals have distorted and distorted reality. Indeed, for example, was Facebook’s Cambridge Analytica case, for which the New York Times accused Facebook of allowing the company to use data from nearly 50 million users of Donald Trump to influence voters’ vote in favor of the Republican nominee. done for in 2016. Recently photographs of black men being defined by algorithms as “apes” or even right-wing politics rather than left-wing politics have recently sparked a great deal of debate on Twitter. ..
So many criticisms that can be algorithmically reevaluated if they more or less control our lives on social networks (recently, Instagram has allowed its users to gain control of their news feed, for example). Offered).
As defended by Aureli Jean (cited above), algorithms are standardized by humans and it is up to us to develop them. will change? will not change?
Only the future will tell us.
4 Handwritten article between Cyril Tiar and Laurian Kifunjuka