Influential algorithms or control over our brains?

Computing power and its influence, also in education

Semantic origin

The word algorithm is said to come from the Latinized name of a Persian mathematician Al-Khwârizmî of the 9th century. The writings of this mathematician in particular the "Treatise on the Indian Numbering System" is said to have helped introduce mathematics to the West.

Definition of an algorithm

"An algorithm is the description of a sequence of steps to obtain a result from elements provided as input"

In the sequence of steps to be performed, there are 3 main features of the algorithm:

a specific output is targeted;
steps follow each other continuously;
the result of the algorithm is known only after the whole process is completed.

To read this definition, following the recipe for making a pizza to the letter is to enter algorithmic thinking. But with computing power and the simultaneous volume of possible iterations, the notion of algorithm takes on a whole new scope.

The influence of algorithms

Often the term algorithm is associated with computer programs, when it was first involved in the construction of mathematical language as with Euclid's algorithm devised 300 A.D.C. to find the greatest common divisors of two positive numbers or integers.

There is a variety of algorithms that focus on dealing with graphical, geometric, data compression, cryptography, decision trees, optimization, etc. They allow considerable time savings in the way of dealing with the specific issues they touch.

Every moment of everyday life is subject to algorithms so much has the power of computation penetrated all our objects and all our networks.

One of the most famous algorithms is the Google's PageRank created in 1996; this algorithm indexes pages according to their relevance, defined according to several criteria including popularity, and eventually became a universal tool for ranking the world. Its usage statistics are edifying because 130,000 billion pages are indexed by Google, 20 billion sites are visited every day. 80,000 requests every second, that is to say 6.9 billion per day. 15% of requests are new requests. With 60% of the search engine market share, Google has imposed itself on our way of thinking well before its competitors. Besides, who uses Qwant? Who remembers Alta Vista?

Its success is due to its ability to evaluate the quality and number of links to a web page to get an estimate of the importance of the website in question. The more the site is cited the more popular it is, the more quality information it is presumed to contain.

The challenges of algorithms

One of the limitations/dangers of algorithms is that they place the web user in a "filter bubble" since the algorithm, for example that of Facebook, customizes access to information without the user's knowledge. The things he sees and the things he doesn't see are meticulously sorted by the computer tool according to the preferences and clicks recorded during each visit. Their browsing history is captured by cookies. In doing so, the Internet user is locked in a space supposedly open to the world but ultimately closed to his inclinations and attempts to influence the owners of algorithms.

In addition to this confinement, a suspicion of uncontrolled complexity was revealed by the "Facebook Files", hundreds of internal documents at Facebook were copied by Frances Haugen, a specialist in algorithms and transmitted to the U.S. regulator. These files show that Facebook's algorithm has become opaque, and even less and less controllable, by Facebook employees. Unintended effects are noted.

Ethical issues are also noted as algorithms tend to carry societal biases with them. For example a research showed how the recruitment platform, from Amazon, discriminated against female applicants because female applicants mentioned fewer keywords than male applicants. The algorithm then extends gender biases by favoring men in the recruiters' search.

Algorithms are therefore subject to the biases of the designers whether these are conscious or unconscious. The result is that decisions entrusted to these programs reinforce the balances in place, even when these are discriminatory.

Knowledge algorithms

In the field of knowledge and on "online learning platforms" algorithms claim to guide our learning and to pace the exercises that are good to follow according to our "level". These algorithms are particularly perverse because they diminish the ability to self-direct one's learning, an essential condition for the sustainability of motivation. The learner is also captured in a "filter bubble" and distanced from his or her own analysis of himself or herself.

The professions are subject to proposals from consulting firms that claim to be able to automate tasks and alleviate the decision-making of experts whose knowledge is carried by memorizing streams of information.

There remains a need to examine the biases of designers on their learning beliefs to avoid reverting back to behavioral control practices. Perhaps mixed teams of experienced pedagogues and computer developers would limit the risks and find novel solutions.

Sources

Cnil algorithm definition - https://www.cnil.fr/en/definition/algorithm#:

Genius.inc 8 most important algorithms in history
https://www.genie-inc.com/8-algorithmes-plus-importants-histoire/

Digital news. All the stats you need to know in 2020
https://www.109c.fr/actualite-du-digital/chiffres-google-toutes-les-statistiques-a-connaitre-en-2020

Wikipedia List of algorithms https://fr.m.wikipedia.org/wiki/Liste_d%27algorithmes

Edrawsoft. Algorithm definition https://www.edrawsoft.com/fr/algorithm-definition.html

The World. How Facebook's algorithm is getting out of control of its creators
https://www.lemonde.fr/pixels/article/2021/10/26/comment-l-algorithme-de-facebook-echappe-au-controle-de-ses-createurs_6099888_4408996.html

Daily geek show https://dailygeekshow.com/internet-informatique-algorithme-monde/

Wikipedia. Filter bubble https://fr.m.wikipedia.org/wiki/Bulle_de_filtres

Ljallamion. Muhammad Ibn al-Khuwrizm? aka Al-Khwârizmî or Al-Khwarizmin https://www.ljallamion.fr/spip.php?article8210

The Echoes. The ethical issues of algorithms
https://www.lesechos.fr/idees-debats/leadership-management/les-enjeux-ethiques-des-algorithmes-1245089#